require(fBasics)
require(hnp)
library(MVN)Dados
dados = read.table("Dados_Pontos1.txt", header = TRUE)Descrição do Banco de Dados
O estudo foi realizado com todo o estado do Piauí dos elementos de terras rara. Onde, o levantamento amostral foi realizado no ano passado (2018). O bando de dados contém 243 obsrvações e 32 variáveis, sendo então, 3 delas variáveis respostas: areia; argila e silte. Que tem como objetivo de verificar se há uma relação das variáveis Elementos terras raras como as variáveis areia, argila e silte.
Visualizando as 6 primeiras linhas da base
head(dados) Areia Argila Silte La Ce Pr Nd Sm Eu Gd Tb Dy Er
1 49.75 32.48 17.77 2.33 8.90 1.08 3.73 1.48 0.23 0.05 0.58 0.00 0.60
2 70.03 27.89 2.08 1.28 3.95 0.50 1.75 0.55 0.10 0.05 0.15 0.00 0.20
3 68.90 28.31 2.79 3.20 6.78 1.18 3.13 0.83 0.13 0.40 0.15 0.30 0.23
4 76.14 22.08 1.79 84.20 165.78 34.10 57.13 10.55 0.63 6.28 1.30 3.35 1.23
5 74.77 19.98 5.25 43.63 145.03 24.25 30.63 5.83 0.70 3.60 1.03 2.28 1.18
6 70.00 27.21 2.79 87.45 128.90 32.70 72.45 13.25 1.25 7.20 1.58 4.25 2.20
Yb Lu ETRLs ETRPs ETRs ETRLs_ETRPs Ca Mg P K Al H_Al
1 0.65 1.10 17.73 2.98 20.70 5.96 0.0 0.27 1.56 24.00 1.4 8.92
2 0.25 0.35 8.13 1.00 9.13 8.13 0.0 0.20 1.04 18.48 0.8 6.73
3 0.25 0.20 15.23 1.53 16.75 9.98 0.0 0.00 2.53 39.63 1.2 5.31
4 0.65 0.20 352.38 13.00 365.38 27.11 0.5 0.15 2.14 105.83 1.0 1.52
5 0.93 0.40 250.05 9.40 259.45 26.60 1.7 1.12 3.96 198.70 0.4 2.05
6 1.60 0.50 336.00 17.33 353.33 19.39 0.8 0.47 5.39 226.29 0.2 1.61
SB V T m T_2 Carbono pH
1 24.27 73.13 33.19 2.80 33.19 1.50 4.22
2 18.68 73.52 25.41 2.10 25.41 0.91 4.40
3 39.63 88.19 44.94 1.49 44.94 0.83 3.80
4 106.51 98.60 108.03 0.47 108.03 0.40 5.16
5 201.54 98.99 203.60 0.10 203.60 0.46 5.45
6 227.59 99.30 229.21 0.04 229.21 0.53 5.43attach(dados)Resumo Descritivo
a<-round(basicStats(dados),2)| Areia | Argila | Silte | La | Ce | Pr | Nd | Sm | Eu | Gd | Tb | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| nobs | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 |
| NAs | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| Minimum | 18.89 | 0.10 | 0.03 | 0.65 | 1.13 | 0.00 | 0.05 | 0.10 | 0.00 | 0.00 | 0.00 |
| Maximum | 98.97 | 50.80 | 60.86 | 227.40 | 377.10 | 46.10 | 155.00 | 33.00 | 5.80 | 29.15 | 5.10 |
| 1. Quartile | 63.71 | 10.95 | 1.98 | 2.98 | 7.72 | 0.94 | 2.65 | 0.63 | 0.10 | 0.10 | 0.05 |
| 3. Quartile | 83.61 | 22.26 | 15.02 | 25.23 | 56.08 | 7.75 | 20.15 | 4.22 | 0.64 | 1.70 | 0.55 |
| Mean | 72.32 | 17.24 | 10.44 | 23.19 | 48.10 | 6.17 | 19.10 | 3.64 | 0.68 | 1.81 | 0.49 |
| Median | 75.74 | 16.40 | 5.03 | 9.90 | 24.38 | 2.78 | 8.28 | 1.60 | 0.25 | 0.50 | 0.20 |
| Sum | 17574.95 | 4189.16 | 2536.10 | 5635.82 | 11689.06 | 1499.84 | 4640.25 | 885.53 | 164.37 | 439.64 | 118.30 |
| SE Mean | 0.97 | 0.59 | 0.79 | 2.24 | 4.22 | 0.54 | 1.78 | 0.33 | 0.07 | 0.22 | 0.05 |
| LCL Mean | 70.41 | 16.08 | 8.88 | 18.78 | 39.79 | 5.12 | 15.59 | 2.99 | 0.54 | 1.37 | 0.39 |
| UCL Mean | 74.24 | 18.40 | 11.99 | 27.60 | 56.41 | 7.23 | 22.60 | 4.30 | 0.81 | 2.25 | 0.58 |
| Variance | 229.05 | 84.42 | 151.03 | 1218.03 | 4323.84 | 69.87 | 769.80 | 27.11 | 1.12 | 12.27 | 0.56 |
| Stdev | 15.13 | 9.19 | 12.29 | 34.90 | 65.76 | 8.36 | 27.75 | 5.21 | 1.06 | 3.50 | 0.75 |
| Skewness | -0.91 | 0.88 | 1.70 | 2.73 | 2.39 | 2.23 | 2.48 | 2.62 | 2.69 | 3.86 | 2.72 |
| Kurtosis | 0.50 | 1.32 | 2.44 | 8.62 | 6.34 | 5.05 | 6.54 | 7.83 | 7.64 | 19.62 | 8.90 |
| Dy | Er | Yb | Lu | ETRLs | ETRPs | ETRs | ETRLs_ETRPs | Ca | Mg | |
|---|---|---|---|---|---|---|---|---|---|---|
| nobs | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 |
| NAs | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| Minimum | 0.00 | 0.00 | 0.00 | 0.00 | 2.45 | 0.05 | 2.58 | 4.77 | 0.00 | 0.00 |
| Maximum | 51.30 | 14.35 | 11.00 | 2.35 | 793.25 | 95.15 | 813.20 | 64.23 | 22.90 | 33.10 |
| 1. Quartile | 0.20 | 0.15 | 0.10 | 0.00 | 15.14 | 1.01 | 16.16 | 11.12 | 0.00 | 0.10 |
| 3. Quartile | 2.40 | 1.05 | 0.80 | 0.32 | 112.25 | 7.28 | 117.28 | 24.35 | 1.90 | 1.50 |
| Mean | 2.97 | 1.08 | 0.87 | 0.25 | 100.88 | 7.46 | 108.34 | 18.54 | 1.64 | 2.00 |
| Median | 0.90 | 0.35 | 0.30 | 0.10 | 48.23 | 2.55 | 51.75 | 16.17 | 0.40 | 0.47 |
| Sum | 722.44 | 261.26 | 210.64 | 61.93 | 24513.46 | 1813.71 | 26326.91 | 4505.57 | 399.50 | 487.08 |
| SE Mean | 0.38 | 0.13 | 0.10 | 0.02 | 8.97 | 0.85 | 9.66 | 0.67 | 0.19 | 0.26 |
| LCL Mean | 2.23 | 0.83 | 0.66 | 0.21 | 83.21 | 5.80 | 89.31 | 17.23 | 1.26 | 1.50 |
| UCL Mean | 3.71 | 1.32 | 1.07 | 0.30 | 118.55 | 9.13 | 127.37 | 19.85 | 2.02 | 2.51 |
| Variance | 34.48 | 3.83 | 2.63 | 0.15 | 19553.10 | 173.93 | 22686.23 | 107.69 | 9.03 | 16.02 |
| Stdev | 5.87 | 1.96 | 1.62 | 0.39 | 139.83 | 13.19 | 150.62 | 10.38 | 3.00 | 4.00 |
| Skewness | 4.04 | 3.47 | 3.57 | 2.55 | 2.32 | 3.52 | 2.31 | 1.32 | 3.66 | 3.53 |
| Kurtosis | 22.40 | 14.08 | 14.20 | 7.34 | 5.68 | 15.20 | 5.48 | 2.39 | 18.25 | 17.13 |
| P | K | Al | H_Al | SB | V | T | m | T_2 | Carbono | pH | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| nobs | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 | 243.00 |
| NAs | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| Minimum | 0.00 | 0.60 | 0.00 | 0.40 | 6.48 | 37.35 | 8.87 | 0.00 | 8.87 | 0.01 | 3.11 |
| Maximum | 330.99 | 364.21 | 4.80 | 17.50 | 524.68 | 99.72 | 530.89 | 9.70 | 530.89 | 2.50 | 8.57 |
| 1. Quartile | 2.66 | 17.83 | 0.20 | 1.28 | 24.94 | 89.62 | 28.41 | 0.11 | 28.41 | 0.38 | 3.80 |
| 3. Quartile | 7.82 | 77.57 | 1.00 | 5.06 | 107.53 | 98.12 | 109.56 | 1.61 | 109.56 | 1.28 | 5.42 |
| Mean | 18.02 | 54.52 | 0.70 | 3.67 | 73.47 | 91.62 | 77.14 | 1.14 | 77.14 | 0.87 | 4.72 |
| Median | 4.41 | 35.54 | 0.60 | 2.77 | 48.76 | 95.17 | 53.13 | 0.55 | 53.13 | 0.64 | 4.55 |
| Sum | 4379.56 | 13248.30 | 170.00 | 892.03 | 17853.47 | 22263.58 | 18745.66 | 277.96 | 18745.66 | 210.97 | 1146.59 |
| SE Mean | 2.42 | 3.43 | 0.04 | 0.20 | 4.53 | 0.65 | 4.53 | 0.10 | 4.53 | 0.04 | 0.07 |
| LCL Mean | 13.25 | 47.75 | 0.62 | 3.27 | 64.56 | 90.34 | 68.22 | 0.95 | 68.22 | 0.78 | 4.58 |
| UCL Mean | 22.79 | 61.29 | 0.78 | 4.07 | 82.39 | 92.89 | 86.06 | 1.34 | 86.06 | 0.95 | 4.86 |
| Variance | 1423.84 | 2866.93 | 0.36 | 9.97 | 4976.24 | 101.83 | 4984.49 | 2.36 | 4984.49 | 0.47 | 1.17 |
| Stdev | 37.73 | 53.54 | 0.60 | 3.16 | 70.54 | 10.09 | 70.60 | 1.54 | 70.60 | 0.68 | 1.08 |
| Skewness | 3.99 | 2.06 | 1.87 | 1.71 | 2.57 | -2.57 | 2.60 | 2.63 | 2.60 | 0.93 | 0.82 |
| Kurtosis | 21.92 | 6.61 | 8.80 | 3.55 | 10.29 | 8.20 | 10.53 | 9.53 | 10.53 | -0.31 | 0.32 |
Gráfico para verificar a dispersão dos dados
boxplot(dados, las = 2)
Através da estatística descritiva, vista anteriormente, temos que, de um total de 243 observações,não houveram respostas em branco para nenhuma das variáveis. Além disso, podemos notar que, há uma grande variabilidade na variável “Ce”, “ETRLs”,“ETRs”,“K”, “SB”, “T”, e “T_2” e, consequentemente, um maior desvio.
Teste de Normalidade de Anderson-Darling.
dados1 = dados[,-30,-31]
result1 = mvn(data=dados1, univariateTest = "AD", desc= F);result1$multivariateNormality
Test HZ p value MVN
1 Henze-Zirkler 972 0 NO
$univariateNormality
Test Variable Statistic p value Normality
1 Anderson-Darling Areia 4.4664 <0.001 NO
2 Anderson-Darling Argila 1.9859 <0.001 NO
3 Anderson-Darling Silte 19.2242 <0.001 NO
4 Anderson-Darling La 28.3476 <0.001 NO
5 Anderson-Darling Ce 24.8163 <0.001 NO
6 Anderson-Darling Pr 24.6587 <0.001 NO
7 Anderson-Darling Nd 27.6715 <0.001 NO
8 Anderson-Darling Sm 27.4131 <0.001 NO
9 Anderson-Darling Eu 32.3354 <0.001 NO
10 Anderson-Darling Gd 35.8790 <0.001 NO
11 Anderson-Darling Tb 27.4585 <0.001 NO
12 Anderson-Darling Dy 38.3064 <0.001 NO
13 Anderson-Darling Er 37.9690 <0.001 NO
14 Anderson-Darling Yb 39.3324 <0.001 NO
15 Anderson-Darling Lu 24.4065 <0.001 NO
16 Anderson-Darling ETRLs 25.8977 <0.001 NO
17 Anderson-Darling ETRPs 35.4924 <0.001 NO
18 Anderson-Darling ETRs 26.2071 <0.001 NO
19 Anderson-Darling ETRLs_ETRPs 5.4007 <0.001 NO
20 Anderson-Darling Ca 30.4671 <0.001 NO
21 Anderson-Darling Mg 40.5301 <0.001 NO
22 Anderson-Darling P 49.2275 <0.001 NO
23 Anderson-Darling K 11.8767 <0.001 NO
24 Anderson-Darling Al 4.4747 <0.001 NO
25 Anderson-Darling H_Al 9.4129 <0.001 NO
26 Anderson-Darling SB 13.1057 <0.001 NO
27 Anderson-Darling V 19.3385 <0.001 NO
28 Anderson-Darling T 13.0637 <0.001 NO
29 Anderson-Darling m 17.6742 <0.001 NO
30 Anderson-Darling Carbono 10.7011 <0.001 NO
31 Anderson-Darling pH 3.2873 <0.001 NO De acordo com o teste de Anderson-Darling, observamos que as variáveis não seguem normalidade, pois os valores de p (p-value) contidos acima são menores que o nível de significância (P=0.05).
Correlação de Spearman
Variável Areia
cor.test(dados$Areia,dados$Yb, method = "spearman")
Spearman's rank correlation rho
data: dados$Areia and dados$Yb
S = 4002220, p-value < 2.2e-16
alternative hypothesis: true rho is not equal to 0
sample estimates:
rho
-0.6735577 cor.test(dados$Areia,dados$ETRLs_ETRPs, method = "spearman")
Spearman's rank correlation rho
data: dados$Areia and dados$ETRLs_ETRPs
S = 1666806, p-value = 1.489e-06
alternative hypothesis: true rho is not equal to 0
sample estimates:
rho
0.3030126 Variável Argila
cor.test(dados$Argila,dados$ETRLs_ETRPs, method = "spearman")
Spearman's rank correlation rho
data: dados$Argila and dados$ETRLs_ETRPs
S = 3324672, p-value = 2.91e-10
alternative hypothesis: true rho is not equal to 0
sample estimates:
rho
-0.3902362 cor.test(dados$Argila,dados$Lu, method = "spearman")
Spearman's rank correlation rho
data: dados$Argila and dados$Lu
S = 1125359, p-value < 2.2e-16
alternative hypothesis: true rho is not equal to 0
sample estimates:
rho
0.5294228 Variável Silte
cor.test(dados$Silte,dados$m, method = "spearman")
Spearman's rank correlation rho
data: dados$Silte and dados$m
S = 3113370, p-value = 1.636e-06
alternative hypothesis: true rho is not equal to 0
sample estimates:
rho
-0.3018785 cor.test(dados$Silte,dados$Ca, method = "spearman")
Spearman's rank correlation rho
data: dados$Silte and dados$Ca
S = 1125132, p-value < 2.2e-16
alternative hypothesis: true rho is not equal to 0
sample estimates:
rho
0.5295175 Após observarmos a relação de cada variável explicativa com as variáveis respostas “Areia”,“Argila” e “Silte”, temos que, para variável resp. Areia há uma maior relação (proporcional como inversamente proporcional) entre as variáveis “Yb” e “ETRLs_ETRPS”. Onde que para a Var “Yb” teve uma porcentagem de 67.36% com coeficiente negativo. Ou seja, quando uma aumenta a outra tende a diminuir. Já para a var “ETRLs_ETRPS”, tivemos uma porcetagem de 30.30% pois ambas tendem aumentar em conjunto, tornando o coeficiente positivo.
Para a segunda Variável resposta “Argila”, temos que, a variável “ETRLs_ETRPS”teve uma porcentagem de 39.02% com coeficiente negativo. Ou seja, quando uma aumenta a outra tende a diminuir. Já para a var “Lu”, tivemos uma porcetagem de 52.94%, pois ambas tendem aumentar em conjunto, tornando o coeficiente positivo.
Para a terceira Variável resposta “Silte”, temos que, a variável “m”teve uma porcentagem de 30.19% com coeficiente negativo. Ou seja, quando uma aumenta a outra tende a diminuir. Já para a var “Ca”, tivemos uma porcetagem de 52.95%, pois ambas tendem aumentar em conjunto, tornando o coeficiente positivo.
Modelo Linear Clássico
A Análise de Regressão Linear era considerada a principal técnica de modelagem estatística até meados do século XX. Seu objetivo principal é analisar a relação entre uma variável resposta e uma ou mais variáveis explicativas, e que a variável resposta segue a distribuição Normal.
modelo1 = lm (Areia~., data = dados[,c(-2,-3,-30)])
summary(modelo1)
Call:
lm(formula = Areia ~ ., data = dados[, c(-2, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-39.731 -4.231 0.825 5.894 19.910
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 103.66341 18.25581 5.678 4.4e-08 ***
La -19.83594 112.84309 -0.176 0.860630
Ce -19.85841 112.85038 -0.176 0.860483
Pr -19.76379 112.85565 -0.175 0.861147
Nd -19.80935 112.85912 -0.176 0.860835
Sm -21.60076 112.93820 -0.191 0.848502
Eu -22.39491 112.67152 -0.199 0.842637
Gd -230.99513 154.07867 -1.499 0.135295
Tb -216.34141 154.46660 -1.401 0.162791
Dy -232.25570 154.12888 -1.507 0.133313
Er -240.37485 154.03902 -1.560 0.120124
Yb -224.35114 154.49561 -1.452 0.147924
Lu -255.97637 153.82713 -1.664 0.097566 .
ETRLs -198.94740 246.30840 -0.808 0.420151
ETRPs 13.86536 260.07144 0.053 0.957532
ETRs 218.75513 208.70886 1.048 0.295759
ETRLs_ETRPs 0.13653 0.08283 1.648 0.100753
Ca -0.70461 0.49058 -1.436 0.152385
Mg -1.08510 0.68567 -1.583 0.115002
P 0.09008 0.05032 1.790 0.074825 .
K -0.05775 0.02316 -2.493 0.013414 *
Al -6.32130 1.71870 -3.678 0.000297 ***
H_Al 32.57907 132.50110 0.246 0.806013
SB 33.37234 132.45778 0.252 0.801324
V -0.12323 0.17372 -0.709 0.478879
T -33.35473 132.45686 -0.252 0.801425
m 0.73399 0.98434 0.746 0.456686
Carbono -2.79537 1.07677 -2.596 0.010083 *
pH -0.63607 1.04707 -0.607 0.544179
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.371 on 214 degrees of freedom
Multiple R-squared: 0.661, Adjusted R-squared: 0.6166
F-statistic: 14.9 on 28 and 214 DF, p-value: < 2.2e-16Seleção das Variáveis
#step(modelo1)
mod1 = lm(formula = Areia ~ Sm + Gd + Tb + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + Mg + P + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1)
Call:
lm(formula = Areia ~ Sm + Gd + Tb + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + Mg + P + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-39.596 -3.852 0.831 6.145 19.986
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 87.93826 2.15157 40.872 < 2e-16 ***
Sm -2.48992 0.84130 -2.960 0.00341 **
Gd -236.15843 116.41632 -2.029 0.04369 *
Tb -219.43082 116.57770 -1.882 0.06110 .
Dy -237.97558 116.46906 -2.043 0.04220 *
Er -245.90849 116.51715 -2.110 0.03593 *
Yb -229.84248 116.76444 -1.968 0.05026 .
Lu -262.24437 116.18707 -2.257 0.02497 *
ETRLs -238.17413 116.48001 -2.045 0.04205 *
ETRs 238.14089 116.47961 2.044 0.04208 *
ETRLs_ETRPs 0.16024 0.07772 2.062 0.04039 *
Ca -0.83602 0.46303 -1.806 0.07234 .
Mg -1.14033 0.65419 -1.743 0.08269 .
P 0.09165 0.04778 1.918 0.05637 .
K -0.05845 0.02076 -2.815 0.00531 **
Al -6.68083 1.41860 -4.709 4.37e-06 ***
SB 0.54727 0.23421 2.337 0.02034 *
T -0.52810 0.23102 -2.286 0.02320 *
m 1.33901 0.63110 2.122 0.03497 *
Carbono -2.79998 1.03277 -2.711 0.00723 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.231 on 223 degrees of freedom
Multiple R-squared: 0.6572, Adjusted R-squared: 0.628
F-statistic: 22.5 on 19 and 223 DF, p-value: < 2.2e-16mod1.2 = lm(formula = Areia ~ Sm + Gd + Tb + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + P + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1.2)
Call:
lm(formula = Areia ~ Sm + Gd + Tb + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + P + K + Al + SB + T + m + Carbono,
data = dados[, c(-2, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-39.412 -3.914 1.172 6.040 18.706
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 88.13851 2.15825 40.838 < 2e-16 ***
Sm -2.30839 0.83862 -2.753 0.00640 **
Gd -220.36313 116.59002 -1.890 0.06004 .
Tb -204.88224 116.80642 -1.754 0.08079 .
Dy -222.26978 116.64717 -1.905 0.05800 .
Er -230.41881 116.70522 -1.974 0.04957 *
Yb -215.13621 116.98794 -1.839 0.06724 .
Lu -246.07463 116.34195 -2.115 0.03553 *
ETRLs -222.59311 116.66377 -1.908 0.05767 .
ETRs 222.55601 116.66319 1.908 0.05771 .
ETRLs_ETRPs 0.16257 0.07806 2.083 0.03842 *
Ca -1.41714 0.32280 -4.390 1.75e-05 ***
P 0.02010 0.02457 0.818 0.41409
K -0.05778 0.02085 -2.771 0.00606 **
Al -7.02843 1.41089 -4.982 1.26e-06 ***
SB 0.52109 0.23479 2.219 0.02746 *
T -0.50496 0.23169 -2.179 0.03034 *
m 1.34627 0.63395 2.124 0.03480 *
Carbono -2.55737 1.02800 -2.488 0.01359 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.273 on 224 degrees of freedom
Multiple R-squared: 0.6525, Adjusted R-squared: 0.6246
F-statistic: 23.37 on 18 and 224 DF, p-value: < 2.2e-16mod1.3 = lm(formula = Areia ~ Sm + Gd + Tb + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1.3)
Call:
lm(formula = Areia ~ Sm + Gd + Tb + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + K + Al + SB + T + m + Carbono,
data = dados[, c(-2, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-39.440 -3.885 1.101 6.068 17.218
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 88.23545 2.15342 40.975 < 2e-16 ***
Sm -2.25487 0.83545 -2.699 0.00748 **
Gd -224.73733 116.38183 -1.931 0.05474 .
Tb -209.53890 116.58198 -1.797 0.07362 .
Dy -226.55621 116.44387 -1.946 0.05295 .
Er -234.88952 116.49160 -2.016 0.04495 *
Yb -219.49847 116.78054 -1.880 0.06146 .
Lu -250.58687 116.12580 -2.158 0.03200 *
ETRLs -226.94427 116.45689 -1.949 0.05257 .
ETRs 226.90743 116.45630 1.948 0.05261 .
ETRLs_ETRPs 0.15655 0.07766 2.016 0.04499 *
Ca -1.31517 0.29756 -4.420 1.54e-05 ***
K -0.05851 0.02082 -2.810 0.00539 **
Al -7.26665 1.37951 -5.268 3.23e-07 ***
SB 0.52244 0.23461 2.227 0.02695 *
T -0.50205 0.23149 -2.169 0.03115 *
m 1.45258 0.62004 2.343 0.02002 *
Carbono -2.62821 1.02359 -2.568 0.01089 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.266 on 225 degrees of freedom
Multiple R-squared: 0.6515, Adjusted R-squared: 0.6251
F-statistic: 24.74 on 17 and 225 DF, p-value: < 2.2e-16mod1.4 = lm(formula = Areia ~ Sm + Gd + Dy + Er + Yb + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1.4)
Call:
lm(formula = Areia ~ Sm + Gd + Dy + Er + Yb + Lu + ETRLs + ETRs +
ETRLs_ETRPs + Ca + K + Al + SB + T + m + Carbono, data = dados[,
c(-2, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-38.693 -3.977 0.614 6.108 17.981
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 87.84755 2.15312 40.800 < 2e-16 ***
Sm -2.22866 0.83943 -2.655 0.008496 **
Gd -15.85409 6.21878 -2.549 0.011454 *
Dy -17.55861 6.19031 -2.836 0.004976 **
Er -26.21133 9.55268 -2.744 0.006559 **
Yb -9.81093 5.22623 -1.877 0.061772 .
Lu -42.94690 11.85076 -3.624 0.000358 ***
ETRLs -17.92242 6.18168 -2.899 0.004109 **
ETRs 17.88682 6.18351 2.893 0.004193 **
ETRLs_ETRPs 0.15089 0.07798 1.935 0.054232 .
Ca -1.25401 0.29706 -4.221 3.52e-05 ***
K -0.06740 0.02032 -3.316 0.001062 **
Al -7.33482 1.38578 -5.293 2.84e-07 ***
SB 0.47917 0.23452 2.043 0.042194 *
T -0.45507 0.23114 -1.969 0.050203 .
m 1.55266 0.62058 2.502 0.013059 *
Carbono -2.97639 1.01004 -2.947 0.003547 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.312 on 226 degrees of freedom
Multiple R-squared: 0.6465, Adjusted R-squared: 0.6214
F-statistic: 25.83 on 16 and 226 DF, p-value: < 2.2e-16mod1.5 = lm(formula = Areia ~ Sm + Gd + Dy + Er + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1.5)
Call:
lm(formula = Areia ~ Sm + Gd + Dy + Er + Lu + ETRLs + ETRs +
ETRLs_ETRPs + Ca + K + Al + SB + T + m + Carbono, data = dados[,
c(-2, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-38.586 -3.954 0.646 6.209 18.693
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 87.80865 2.16496 40.559 < 2e-16 ***
Sm -1.34872 0.70023 -1.926 0.05534 .
Gd -7.95883 4.60631 -1.728 0.08538 .
Dy -9.65873 4.56507 -2.116 0.03545 *
Er -18.95174 8.78336 -2.158 0.03200 *
Lu -30.01532 9.69648 -3.095 0.00221 **
ETRLs -9.94239 4.51296 -2.203 0.02859 *
ETRs 9.89997 4.51204 2.194 0.02924 *
ETRLs_ETRPs 0.15091 0.07841 1.925 0.05551 .
Ca -1.30733 0.29734 -4.397 1.69e-05 ***
K -0.06717 0.02044 -3.287 0.00117 **
Al -7.29957 1.39333 -5.239 3.68e-07 ***
SB 0.53741 0.23375 2.299 0.02241 *
T -0.51036 0.23053 -2.214 0.02783 *
m 1.45642 0.62188 2.342 0.02005 *
Carbono -3.09901 1.01351 -3.058 0.00250 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.364 on 227 degrees of freedom
Multiple R-squared: 0.6409, Adjusted R-squared: 0.6172
F-statistic: 27.01 on 15 and 227 DF, p-value: < 2.2e-16mod1.6 = lm(formula = Areia ~ Sm + Dy + Er + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1.6)
Call:
lm(formula = Areia ~ Sm + Dy + Er + Lu + ETRLs + ETRs + ETRLs_ETRPs +
Ca + K + Al + SB + T + m + Carbono, data = dados[, c(-2,
-3, -30)])
Residuals:
Min 1Q Median 3Q Max
-38.877 -3.943 0.820 6.186 19.845
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 87.82514 2.17434 40.392 < 2e-16 ***
Sm -1.18911 0.69713 -1.706 0.08942 .
Dy -1.82789 0.54899 -3.330 0.00101 **
Er -4.24155 2.16876 -1.956 0.05172 .
Lu -14.21125 3.23213 -4.397 1.68e-05 ***
ETRLs -2.18966 0.48520 -4.513 1.02e-05 ***
ETRs 2.14921 0.48723 4.411 1.59e-05 ***
ETRLs_ETRPs 0.15129 0.07875 1.921 0.05596 .
Ca -1.36472 0.29676 -4.599 7.05e-06 ***
K -0.06466 0.02047 -3.158 0.00180 **
Al -7.53240 1.39283 -5.408 1.61e-07 ***
SB 0.59606 0.23227 2.566 0.01092 *
T -0.57107 0.22883 -2.496 0.01328 *
m 1.27709 0.61582 2.074 0.03922 *
Carbono -2.85136 1.00769 -2.830 0.00508 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.404 on 228 degrees of freedom
Multiple R-squared: 0.6362, Adjusted R-squared: 0.6139
F-statistic: 28.48 on 14 and 228 DF, p-value: < 2.2e-16mod1.7 = lm(formula = Areia ~ Dy + Er + Lu + ETRLs +
ETRs + ETRLs_ETRPs + Ca + K + Al + SB + T + m +
Carbono, data = dados[, c(-2, -3, -30)])
summary(mod1.7)
Call:
lm(formula = Areia ~ Dy + Er + Lu + ETRLs + ETRs + ETRLs_ETRPs +
Ca + K + Al + SB + T + m + Carbono, data = dados[, c(-2,
-3, -30)])
Residuals:
Min 1Q Median 3Q Max
-38.975 -3.740 1.147 6.214 20.407
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 87.54228 2.17703 40.212 < 2e-16 ***
Dy -1.29470 0.45319 -2.857 0.00467 **
Er -4.45909 2.17401 -2.051 0.04140 *
Lu -14.13713 3.24529 -4.356 2.00e-05 ***
ETRLs -1.75839 0.41584 -4.229 3.40e-05 ***
ETRs 1.69709 0.41052 4.134 5.00e-05 ***
ETRLs_ETRPs 0.15834 0.07897 2.005 0.04613 *
Ca -1.32180 0.29692 -4.452 1.33e-05 ***
K -0.06104 0.02045 -2.985 0.00314 **
Al -7.57780 1.39837 -5.419 1.51e-07 ***
SB 0.64539 0.23142 2.789 0.00574 **
T -0.62477 0.22759 -2.745 0.00653 **
m 1.37981 0.61542 2.242 0.02592 *
Carbono -2.74759 1.01003 -2.720 0.00702 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 9.443 on 229 degrees of freedom
Multiple R-squared: 0.6316, Adjusted R-squared: 0.6107
F-statistic: 30.2 on 13 and 229 DF, p-value: < 2.2e-16Verificando os resíduos do modelo1
par(mfrow = c(2,2))
plot(mod1.7, pch = 20)
Teste de Normalidade para os Resíduos
shapiro.test(mod1.7$residuals)
Shapiro-Wilk normality test
data: mod1.7$residuals
W = 0.95157, p-value = 3.015e-07Ajustando o modelo de regressão para a Variável Argila
modelo2 = lm (Argila ~., data = dados[,c(-1,-3,-30)])
summary(modelo2)
Call:
lm(formula = Argila ~ ., data = dados[, c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-16.782 -3.986 -0.794 3.775 37.059
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 16.58779 13.48167 1.230 0.21990
La -109.43203 83.33315 -1.313 0.19053
Ce -109.40778 83.33852 -1.313 0.19065
Pr -108.44532 83.34242 -1.301 0.19459
Nd -109.85325 83.34498 -1.318 0.18889
Sm -107.05015 83.40338 -1.284 0.20070
Eu -107.81254 83.20644 -1.296 0.19647
Gd -82.21615 113.78508 -0.723 0.47074
Tb -102.29799 114.07156 -0.897 0.37084
Dy -82.38719 113.82216 -0.724 0.46996
Er -81.34950 113.75580 -0.715 0.47531
Yb -84.18427 114.09299 -0.738 0.46141
Lu -60.92413 113.59932 -0.536 0.59230
ETRLs 279.80938 181.89552 1.538 0.12545
ETRPs 252.93216 192.05935 1.317 0.18926
ETRs -170.39230 154.12876 -1.106 0.27018
ETRLs_ETRPs -0.12836 0.06117 -2.098 0.03704 *
Ca -0.54498 0.36229 -1.504 0.13398
Mg 1.11053 0.50636 2.193 0.02937 *
P -0.09816 0.03716 -2.642 0.00886 **
K 0.04521 0.01711 2.643 0.00882 **
Al 2.26859 1.26924 1.787 0.07529 .
H_Al 27.47208 97.85032 0.281 0.77917
SB 26.92040 97.81833 0.275 0.78342
V 0.05335 0.12829 0.416 0.67792
T -26.95525 97.81765 -0.276 0.78315
m -0.34579 0.72692 -0.476 0.63478
Carbono 1.39715 0.79518 1.757 0.08034 .
pH -1.95979 0.77324 -2.534 0.01198 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.92 on 214 degrees of freedom
Multiple R-squared: 0.4984, Adjusted R-squared: 0.4327
F-statistic: 7.593 on 28 and 214 DF, p-value: < 2.2e-16Seleção das varáveis
#step(modelo2)
mod2.1 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Ca + Mg + P + K + Al + H_Al +
T + Carbono + pH, data = dados[, c(-1, -3, -30)])
summary(mod2.1)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Ca + Mg + P + K + Al + H_Al +
T + Carbono + pH, data = dados[, c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-15.018 -4.040 -0.681 3.608 37.575
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 20.15975 3.77532 5.340 2.30e-07 ***
La -1.11553 0.28593 -3.901 0.000127 ***
Ce -1.01072 0.28926 -3.494 0.000574 ***
Nd -1.17675 0.40469 -2.908 0.004009 **
Gd -18.52764 4.29745 -4.311 2.44e-05 ***
Tb -36.62146 8.40426 -4.357 2.01e-05 ***
Dy -18.58731 4.44526 -4.181 4.17e-05 ***
Er -16.27223 3.96472 -4.104 5.70e-05 ***
Yb -20.82824 7.14794 -2.914 0.003935 **
ETRLs 1.01946 0.28181 3.618 0.000368 ***
ETRPs 18.76903 4.45446 4.214 3.66e-05 ***
ETRLs_ETRPs -0.13451 0.05842 -2.303 0.022223 *
Ca -0.52584 0.34971 -1.504 0.134091
Mg 1.07188 0.48735 2.199 0.028880 *
P -0.09771 0.03539 -2.761 0.006253 **
K 0.03679 0.01547 2.377 0.018280 *
Al 2.03602 0.96554 2.109 0.036092 *
H_Al 0.46100 0.17283 2.667 0.008207 **
T -0.02277 0.01372 -1.659 0.098586 .
Carbono 1.32149 0.75402 1.753 0.081054 .
pH -1.75329 0.75148 -2.333 0.020538 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.893 on 222 degrees of freedom
Multiple R-squared: 0.4837, Adjusted R-squared: 0.4372
F-statistic: 10.4 on 20 and 222 DF, p-value: < 2.2e-16mod2.2 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al +
T + Carbono + pH, data = dados[, c(-1, -3, -30)])
summary(mod2.2)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al + T +
Carbono + pH, data = dados[, c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-16.933 -4.034 -0.615 3.475 37.565
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 20.96783 3.74743 5.595 6.43e-08 ***
La -1.08707 0.28611 -3.800 0.000187 ***
Ce -0.98841 0.28970 -3.412 0.000766 ***
Nd -1.15448 0.40556 -2.847 0.004830 **
Gd -17.74728 4.27805 -4.148 4.76e-05 ***
Tb -34.87309 8.34694 -4.178 4.22e-05 ***
Dy -17.69265 4.41770 -4.005 8.45e-05 ***
Er -15.74879 3.96056 -3.976 9.46e-05 ***
Yb -19.39392 7.10401 -2.730 0.006839 **
ETRLs 0.99791 0.28224 3.536 0.000495 ***
ETRPs 17.86668 4.42631 4.036 7.46e-05 ***
ETRLs_ETRPs -0.14068 0.05844 -2.407 0.016882 *
Mg 0.54739 0.34132 1.604 0.110179
P -0.07156 0.03092 -2.315 0.021538 *
K 0.03339 0.01535 2.175 0.030679 *
Al 2.24992 0.95770 2.349 0.019683 *
H_Al 0.45013 0.17316 2.599 0.009960 **
T -0.02124 0.01373 -1.548 0.123113
Carbono 1.06885 0.73714 1.450 0.148466
pH -1.91848 0.74551 -2.573 0.010719 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.912 on 223 degrees of freedom
Multiple R-squared: 0.4784, Adjusted R-squared: 0.434
F-statistic: 10.77 on 19 and 223 DF, p-value: < 2.2e-16mod2.3 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al +
T + pH, data = dados[, c(-1, -3, -30)])
summary(mod2.3)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al + T +
pH, data = dados[, c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.429 -4.016 -1.047 3.600 37.027
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 22.33566 3.63567 6.143 3.65e-09 ***
La -1.05891 0.28615 -3.701 0.000271 ***
Ce -0.96456 0.28994 -3.327 0.001027 **
Nd -1.12084 0.40589 -2.761 0.006232 **
Gd -18.44160 4.26161 -4.327 2.27e-05 ***
Tb -35.41621 8.35903 -4.237 3.31e-05 ***
Dy -18.41387 4.40040 -4.185 4.10e-05 ***
Er -17.11469 3.85635 -4.438 1.42e-05 ***
Yb -19.52134 7.12092 -2.741 0.006612 **
ETRLs 0.97357 0.28243 3.447 0.000677 ***
ETRPs 18.58214 4.40954 4.214 3.64e-05 ***
ETRLs_ETRPs -0.15358 0.05790 -2.653 0.008556 **
Mg 0.55013 0.34215 1.608 0.109275
P -0.07169 0.03099 -2.313 0.021622 *
K 0.03232 0.01537 2.103 0.036602 *
Al 2.25755 0.96004 2.352 0.019564 *
H_Al 0.48566 0.17184 2.826 0.005136 **
T -0.02024 0.01374 -1.473 0.142127
pH -2.01220 0.74453 -2.703 0.007406 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.929 on 224 degrees of freedom
Multiple R-squared: 0.4735, Adjusted R-squared: 0.4312
F-statistic: 11.19 on 18 and 224 DF, p-value: < 2.2e-16mod2.4 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al
+ pH, data = dados[, c(-1, -3, -30)])
summary(mod2.4)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al + pH,
data = dados[, c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.665 -3.930 -0.992 3.473 37.122
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 23.05265 3.61230 6.382 9.87e-10 ***
La -1.07472 0.28669 -3.749 0.000226 ***
Ce -0.96834 0.29068 -3.331 0.001010 **
Nd -1.10380 0.40678 -2.714 0.007172 **
Gd -19.05501 4.25223 -4.481 1.18e-05 ***
Tb -37.11710 8.30040 -4.472 1.23e-05 ***
Dy -18.96600 4.39580 -4.315 2.40e-05 ***
Er -17.54451 3.85528 -4.551 8.75e-06 ***
Yb -20.09464 7.12875 -2.819 0.005250 **
ETRLs 0.97668 0.28316 3.449 0.000671 ***
ETRPs 19.15848 4.40355 4.351 2.06e-05 ***
ETRLs_ETRPs -0.14803 0.05792 -2.556 0.011258 *
Mg 0.52046 0.34244 1.520 0.129956
P -0.07524 0.03098 -2.429 0.015938 *
K 0.01634 0.01092 1.497 0.135867
Al 2.21146 0.96202 2.299 0.022436 *
H_Al 0.42310 0.16694 2.534 0.011945 *
pH -2.21967 0.73298 -3.028 0.002747 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.947 on 225 degrees of freedom
Multiple R-squared: 0.4684, Adjusted R-squared: 0.4283
F-statistic: 11.66 on 17 and 225 DF, p-value: < 2.2e-16mod2.4 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al
+ pH, data = dados[, c(-1, -3, -30)])
summary(mod2.4)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + K + Al + H_Al + pH,
data = dados[, c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.665 -3.930 -0.992 3.473 37.122
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 23.05265 3.61230 6.382 9.87e-10 ***
La -1.07472 0.28669 -3.749 0.000226 ***
Ce -0.96834 0.29068 -3.331 0.001010 **
Nd -1.10380 0.40678 -2.714 0.007172 **
Gd -19.05501 4.25223 -4.481 1.18e-05 ***
Tb -37.11710 8.30040 -4.472 1.23e-05 ***
Dy -18.96600 4.39580 -4.315 2.40e-05 ***
Er -17.54451 3.85528 -4.551 8.75e-06 ***
Yb -20.09464 7.12875 -2.819 0.005250 **
ETRLs 0.97668 0.28316 3.449 0.000671 ***
ETRPs 19.15848 4.40355 4.351 2.06e-05 ***
ETRLs_ETRPs -0.14803 0.05792 -2.556 0.011258 *
Mg 0.52046 0.34244 1.520 0.129956
P -0.07524 0.03098 -2.429 0.015938 *
K 0.01634 0.01092 1.497 0.135867
Al 2.21146 0.96202 2.299 0.022436 *
H_Al 0.42310 0.16694 2.534 0.011945 *
pH -2.21967 0.73298 -3.028 0.002747 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.947 on 225 degrees of freedom
Multiple R-squared: 0.4684, Adjusted R-squared: 0.4283
F-statistic: 11.66 on 17 and 225 DF, p-value: < 2.2e-16mod2.5 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + Al + H_Al
+ pH, data = dados[, c(-1, -3, -30)])
summary(mod2.5)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Mg + P + Al + H_Al + pH, data = dados[,
c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.068 -4.273 -0.850 3.165 37.285
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 23.74624 3.59227 6.610 2.73e-10 ***
La -1.19265 0.27640 -4.315 2.39e-05 ***
Ce -1.07800 0.28207 -3.822 0.000171 ***
Nd -1.27103 0.39220 -3.241 0.001372 **
Gd -18.71785 4.25790 -4.396 1.70e-05 ***
Tb -36.69423 8.31832 -4.411 1.59e-05 ***
Dy -18.71724 4.40469 -4.249 3.14e-05 ***
Er -17.61523 3.86556 -4.557 8.50e-06 ***
Yb -19.60305 7.14069 -2.745 0.006533 **
ETRLs 1.08908 0.27377 3.978 9.36e-05 ***
ETRPs 18.89818 4.41217 4.283 2.73e-05 ***
ETRLs_ETRPs -0.15125 0.05804 -2.606 0.009772 **
Mg 0.64987 0.33226 1.956 0.051707 .
P -0.07789 0.03101 -2.512 0.012711 *
Al 2.21248 0.96465 2.294 0.022735 *
H_Al 0.38242 0.16517 2.315 0.021492 *
pH -2.19910 0.73486 -2.993 0.003074 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 6.966 on 226 degrees of freedom
Multiple R-squared: 0.4631, Adjusted R-squared: 0.4251
F-statistic: 12.19 on 16 and 226 DF, p-value: < 2.2e-16mod2.6 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + P + Al + H_Al
+ pH, data = dados[, c(-1, -3, -30)])
summary(mod2.6)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + P + Al + H_Al + pH, data = dados[,
c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-14.127 -4.433 -0.951 3.758 37.106
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 22.50002 3.55724 6.325 1.33e-09 ***
La -1.18068 0.27805 -4.246 3.17e-05 ***
Ce -1.05084 0.28347 -3.707 0.000264 ***
Nd -1.25346 0.39453 -3.177 0.001694 **
Gd -19.60521 4.25993 -4.602 6.95e-06 ***
Tb -37.41999 8.36160 -4.475 1.21e-05 ***
Dy -19.68121 4.40419 -4.469 1.24e-05 ***
Er -17.73404 3.88906 -4.560 8.37e-06 ***
Yb -20.05208 7.18128 -2.792 0.005680 **
ETRLs 1.06991 0.27529 3.886 0.000134 ***
ETRPs 19.66815 4.42184 4.448 1.36e-05 ***
ETRLs_ETRPs -0.14623 0.05835 -2.506 0.012902 *
P -0.02584 0.01601 -1.613 0.108063
Al 2.33089 0.96873 2.406 0.016923 *
H_Al 0.35184 0.16545 2.127 0.034530 *
pH -1.95415 0.72861 -2.682 0.007855 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 7.01 on 227 degrees of freedom
Multiple R-squared: 0.454, Adjusted R-squared: 0.418
F-statistic: 12.59 on 15 and 227 DF, p-value: < 2.2e-16mod2.7 = lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Al + H_Al
+ pH, data = dados[, c(-1, -3, -30)])
summary(mod2.7)
Call:
lm(formula = Argila ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
ETRLs + ETRPs + ETRLs_ETRPs + Al + H_Al + pH, data = dados[,
c(-1, -3, -30)])
Residuals:
Min 1Q Median 3Q Max
-14.153 -4.364 -0.779 3.318 37.227
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 23.72681 3.48721 6.804 8.87e-11 ***
La -1.20344 0.27867 -4.319 2.34e-05 ***
Ce -1.06675 0.28430 -3.752 0.000222 ***
Nd -1.26817 0.39581 -3.204 0.001549 **
Gd -19.50523 4.27442 -4.563 8.23e-06 ***
Tb -37.48036 8.39085 -4.467 1.25e-05 ***
Dy -19.59976 4.41935 -4.435 1.43e-05 ***
Er -17.44434 3.89854 -4.475 1.21e-05 ***
Yb -20.01312 7.20643 -2.777 0.005941 **
ETRLs 1.08531 0.27609 3.931 0.000112 ***
ETRPs 19.54766 4.43672 4.406 1.62e-05 ***
ETRLs_ETRPs -0.13489 0.05812 -2.321 0.021182 *
Al 2.43455 0.96998 2.510 0.012772 *
H_Al 0.32985 0.16546 1.994 0.047396 *
pH -2.31732 0.69539 -3.332 0.001005 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 7.034 on 228 degrees of freedom
Multiple R-squared: 0.4478, Adjusted R-squared: 0.4139
F-statistic: 13.21 on 14 and 228 DF, p-value: < 2.2e-16Verificando os resíduos do modelo2.7
par(mfrow = c(2,2))
plot(mod2.7, pch = 20)
### Teste de Normalidade para os Resíduos
shapiro.test(mod2.7$residuals)
Shapiro-Wilk normality test
data: mod2.7$residuals
W = 0.94173, p-value = 2.98e-08Ajustando o modelo de regressão para a Variável “Silte”
modelo3 = lm (Silte ~., data = dados[,c(-1,-2,-30)])
summary(modelo3)
Call:
lm(formula = Silte ~ ., data = dados[, c(-1, -2, -30)])
Residuals:
Min 1Q Median 3Q Max
-16.612 -4.808 -1.064 2.722 35.828
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -2.025e+01 1.626e+01 -1.245 0.21440
La 1.293e+02 1.005e+02 1.286 0.19968
Ce 1.293e+02 1.005e+02 1.286 0.19972
Pr 1.282e+02 1.005e+02 1.276 0.20342
Nd 1.297e+02 1.005e+02 1.290 0.19838
Sm 1.287e+02 1.006e+02 1.279 0.20221
Eu 1.302e+02 1.004e+02 1.298 0.19577
Gd 3.132e+02 1.372e+02 2.282 0.02346 *
Tb 3.186e+02 1.376e+02 2.316 0.02151 *
Dy 3.146e+02 1.373e+02 2.292 0.02289 *
Er 3.217e+02 1.372e+02 2.345 0.01995 *
Yb 3.085e+02 1.376e+02 2.242 0.02599 *
Lu 3.169e+02 1.370e+02 2.313 0.02168 *
ETRLs -8.087e+01 2.194e+02 -0.369 0.71279
ETRPs -2.668e+02 2.317e+02 -1.152 0.25078
ETRs -4.839e+01 1.859e+02 -0.260 0.79488
ETRLs_ETRPs -8.140e-03 7.378e-02 -0.110 0.91226
Ca 1.250e+00 4.370e-01 2.861 0.00464 **
Mg -2.578e-02 6.108e-01 -0.042 0.96637
P 8.095e-03 4.482e-02 0.181 0.85684
K 1.253e-02 2.063e-02 0.607 0.54427
Al 4.052e+00 1.531e+00 2.647 0.00873 **
H_Al -6.007e+01 1.180e+02 -0.509 0.61133
SB -6.031e+01 1.180e+02 -0.511 0.60978
V 6.987e-02 1.547e-01 0.452 0.65204
T 6.032e+01 1.180e+02 0.511 0.60967
m -3.881e-01 8.768e-01 -0.443 0.65846
Carbono 1.398e+00 9.591e-01 1.458 0.14640
pH 2.595e+00 9.327e-01 2.783 0.00587 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.347 on 214 degrees of freedom
Multiple R-squared: 0.5921, Adjusted R-squared: 0.5387
F-statistic: 11.09 on 28 and 214 DF, p-value: < 2.2e-16###Seleção das variáveis
#step(modelo3)
mod3.1 = lm(formula = Silte ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T + Carbono +
pH, data = dados[, c(-1, -2, -30)])
summary(mod3.1)
Call:
lm(formula = Silte ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T + Carbono +
pH, data = dados[, c(-1, -2, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.616 -4.865 -1.092 2.568 37.029
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -13.67491 4.45824 -3.067 0.00243 **
La 160.26332 95.29828 1.682 0.09403 .
Ce 160.28210 95.30109 1.682 0.09400 .
Pr 159.21956 95.30418 1.671 0.09619 .
Nd 160.73507 95.29391 1.687 0.09305 .
Sm 159.47582 95.47148 1.670 0.09624 .
Eu 161.45168 95.07119 1.698 0.09086 .
Gd 332.00187 130.31943 2.548 0.01152 *
Tb 338.20653 130.61864 2.589 0.01025 *
Dy 333.50072 130.35371 2.558 0.01118 *
Er 340.39357 130.29937 2.612 0.00960 **
Yb 327.74039 130.69600 2.508 0.01287 *
Lu 334.26072 130.38316 2.564 0.01101 *
ETRLs -160.24168 95.29881 -1.681 0.09407 .
ETRPs -334.01053 130.37711 -2.562 0.01107 *
Ca 1.27000 0.28174 4.508 1.06e-05 ***
Al 3.62032 1.10914 3.264 0.00127 **
T 0.03358 0.01160 2.894 0.00418 **
Carbono 1.21803 0.86779 1.404 0.16183
pH 2.59736 0.89177 2.913 0.00395 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.22 on 223 degrees of freedom
Multiple R-squared: 0.5878, Adjusted R-squared: 0.5526
F-statistic: 16.73 on 19 and 223 DF, p-value: < 2.2e-16mod3.2 = lm(formula = Silte ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T +
pH, data = dados[, c(-1, -2, -30)])
summary(mod3.2)
Call:
lm(formula = Silte ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T + pH, data = dados[,
c(-1, -2, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.078 -4.723 -1.218 2.362 38.317
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -12.12883 4.32937 -2.802 0.00553 **
La 161.73485 95.49864 1.694 0.09173 .
Ce 161.74265 95.50153 1.694 0.09173 .
Pr 160.63910 95.50495 1.682 0.09396 .
Nd 162.19388 95.49435 1.698 0.09081 .
Sm 160.98270 95.67193 1.683 0.09384 .
Eu 162.97536 95.27062 1.711 0.08853 .
Gd 336.01389 130.56989 2.573 0.01071 *
Tb 342.23220 130.86961 2.615 0.00953 **
Dy 337.47438 130.60486 2.584 0.01040 *
Er 343.67670 130.56017 2.632 0.00907 **
Yb 332.25357 130.93905 2.537 0.01185 *
Lu 339.37606 130.61413 2.598 0.00999 **
ETRLs -161.70339 95.49924 -1.693 0.09180 .
ETRPs -337.98837 130.62825 -2.587 0.01030 *
Ca 1.31661 0.28038 4.696 4.63e-06 ***
Al 3.67889 1.11075 3.312 0.00108 **
T 0.03356 0.01163 2.887 0.00427 **
pH 2.43525 0.88617 2.748 0.00648 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.238 on 224 degrees of freedom
Multiple R-squared: 0.5841, Adjusted R-squared: 0.5507
F-statistic: 17.48 on 18 and 224 DF, p-value: < 2.2e-16mod3.3 = lm(formula = Silte ~ La + Ce + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T +
pH, data = dados[, c(-1, -2, -30)])
summary(mod3.3)
Call:
lm(formula = Silte ~ La + Ce + Nd + Sm + Eu + Gd + Tb + Dy +
Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T + pH, data = dados[,
c(-1, -2, -30)])
Residuals:
Min 1Q Median 3Q Max
-16.883 -4.553 -1.170 2.175 38.213
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -10.11077 4.17667 -2.421 0.01628 *
La 1.10761 0.37734 2.935 0.00368 **
Ce 1.11043 0.36028 3.082 0.00231 **
Nd 1.57470 0.49025 3.212 0.00151 **
Sm 0.07957 1.38984 0.057 0.95440
Eu 2.77083 2.14866 1.290 0.19853
Gd 400.92501 125.24243 3.201 0.00157 **
Tb 409.31223 125.15051 3.271 0.00124 **
Dy 402.19827 125.31369 3.210 0.00152 **
Er 406.42758 125.62396 3.235 0.00140 **
Yb 398.66515 125.35060 3.180 0.00168 **
Lu 403.56130 125.42211 3.218 0.00148 **
ETRLs -1.07495 0.35062 -3.066 0.00244 **
ETRPs -402.72710 125.33553 -3.213 0.00151 **
Ca 1.30569 0.28144 4.639 5.93e-06 ***
Al 3.62278 1.11476 3.250 0.00133 **
T 0.03519 0.01163 3.026 0.00277 **
pH 2.20022 0.87863 2.504 0.01298 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.271 on 225 degrees of freedom
Multiple R-squared: 0.5789, Adjusted R-squared: 0.547
F-statistic: 18.19 on 17 and 225 DF, p-value: < 2.2e-16mod3.4 = lm(formula = Silte ~ La + Ce + Nd + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T +
pH, data = dados[, c(-1, -2, -30)])
summary(mod3.4)
Call:
lm(formula = Silte ~ La + Ce + Nd + Eu + Gd + Tb + Dy + Er +
Yb + Lu + ETRLs + ETRPs + Ca + Al + T + pH, data = dados[,
c(-1, -2, -30)])
Residuals:
Min 1Q Median 3Q Max
-16.885 -4.572 -1.181 2.174 38.238
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -10.11491 4.16683 -2.427 0.01599 *
La 1.09815 0.33847 3.244 0.00136 **
Ce 1.10362 0.33937 3.252 0.00132 **
Nd 1.57742 0.48686 3.240 0.00138 **
Eu 2.79228 2.11106 1.323 0.18727
Gd 400.10087 124.13775 3.223 0.00146 **
Tb 408.59300 124.24349 3.289 0.00117 **
Dy 401.36530 124.19144 3.232 0.00141 **
Er 405.61658 124.54719 3.257 0.00130 **
Yb 397.82298 124.20966 3.203 0.00156 **
Lu 402.64929 124.13172 3.244 0.00136 **
ETRLs -1.06827 0.32988 -3.238 0.00138 **
ETRPs -401.89252 124.21009 -3.236 0.00140 **
Ca 1.30307 0.27709 4.703 4.47e-06 ***
Al 3.62905 1.10691 3.279 0.00121 **
T 0.03536 0.01126 3.140 0.00191 **
pH 2.20239 0.87588 2.514 0.01262 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.253 on 226 degrees of freedom
Multiple R-squared: 0.5789, Adjusted R-squared: 0.549
F-statistic: 19.41 on 16 and 226 DF, p-value: < 2.2e-16mod3.5 = lm(formula = Silte ~ La + Ce + Nd + Gd + Tb +
Dy + Er + Yb + Lu + ETRLs + ETRPs + Ca + Al + T +
pH, data = dados[, c(-1, -2, -30)])
summary(mod3.5)
Call:
lm(formula = Silte ~ La + Ce + Nd + Gd + Tb + Dy + Er + Yb +
Lu + ETRLs + ETRPs + Ca + Al + T + pH, data = dados[, c(-1,
-2, -30)])
Residuals:
Min 1Q Median 3Q Max
-17.191 -4.630 -1.086 2.175 37.699
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -10.24991 4.17245 -2.457 0.014777 *
La 1.00678 0.33189 3.033 0.002699 **
Ce 1.07190 0.33908 3.161 0.001786 **
Nd 1.71168 0.47695 3.589 0.000407 ***
Gd 404.47846 124.29833 3.254 0.001311 **
Tb 412.23128 124.41794 3.313 0.001073 **
Dy 406.02848 124.34617 3.265 0.001262 **
Er 410.60605 124.69540 3.293 0.001150 **
Yb 402.38933 124.36649 3.236 0.001395 **
Lu 407.83935 124.27435 3.282 0.001194 **
ETRLs -1.03877 0.32967 -3.151 0.001847 **
ETRPs -406.48042 124.36646 -3.268 0.001249 **
Ca 1.40503 0.26659 5.270 3.16e-07 ***
Al 3.63853 1.10872 3.282 0.001194 **
T 0.03231 0.01104 2.927 0.003771 **
pH 2.25168 0.87653 2.569 0.010845 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.266 on 227 degrees of freedom
Multiple R-squared: 0.5756, Adjusted R-squared: 0.5476
F-statistic: 20.52 on 15 and 227 DF, p-value: < 2.2e-16Verificando os resíduos do mod3.5
par(mfrow = c(2,2))
plot(mod3.5, pch = 20)
Teste de Normalidade para os Resíduos
shapiro.test(mod3.5$residuals)
Shapiro-Wilk normality test
data: mod3.5$residuals
W = 0.87331, p-value = 2.516e-13Envelope
hnp(mod1.7, print= T, main="mod1.7 Areia")Gaussian model (lm object) 
hnp(mod2.7, print= T, main="mod2.7 Argila")Gaussian model (lm object) 
hnp(mod3.5, print= T, main= "mod3.5 Silte")Gaussian model (lm object) 
Modelo Linear Generalizado
Como a variável resposta “Areia”, “Argila e”Silte é contínua e assimétrica serão testadas as distribuições Gamma, Normal Inversa e Normal para a modelagem.
Distribuição Gamma para Variável Areia
mlg4 = glm(Areia~., data = dados[,c(-2,-3,-30)], family = Gamma(link="inverse"))
summary(mlg4)
Call:
glm(formula = Areia ~ ., family = Gamma(link = "inverse"), data = dados[,
c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.62212 -0.06714 0.00381 0.08156 0.39300
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.034e-02 3.722e-03 2.777 0.00597 **
La 1.540e-02 2.314e-02 0.665 0.50651
Ce 1.542e-02 2.314e-02 0.666 0.50605
Pr 1.545e-02 2.315e-02 0.668 0.50514
Nd 1.545e-02 2.315e-02 0.668 0.50505
Sm 1.559e-02 2.316e-02 0.673 0.50147
Eu 1.638e-02 2.311e-02 0.709 0.47904
Gd 2.720e-02 3.320e-02 0.819 0.41345
Tb 2.519e-02 3.329e-02 0.757 0.45013
Dy 2.768e-02 3.321e-02 0.834 0.40547
Er 2.968e-02 3.320e-02 0.894 0.37234
Yb 2.624e-02 3.331e-02 0.788 0.43167
Lu 3.267e-02 3.313e-02 0.986 0.32518
ETRLs 3.362e-02 4.991e-02 0.673 0.50136
ETRPs 2.122e-02 5.410e-02 0.392 0.69523
ETRs -4.903e-02 4.298e-02 -1.141 0.25525
ETRLs_ETRPs -9.495e-06 1.686e-05 -0.563 0.57387
Ca 3.301e-04 1.589e-04 2.077 0.03901 *
Mg 4.500e-04 2.078e-04 2.166 0.03143 *
P -3.989e-05 1.461e-05 -2.730 0.00687 **
K 7.019e-06 5.826e-06 1.205 0.22963
Al 1.647e-03 4.205e-04 3.917 0.00012 ***
H_Al -1.653e-02 2.670e-02 -0.619 0.53658
SB -1.660e-02 2.670e-02 -0.622 0.53467
V -1.118e-06 3.536e-05 -0.032 0.97480
T 1.660e-02 2.670e-02 0.622 0.53475
m -3.050e-04 2.082e-04 -1.465 0.14443
Carbono 4.016e-04 2.239e-04 1.793 0.07434 .
pH 6.192e-05 2.164e-04 0.286 0.77500
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.01978286)
Null deviance: 13.419 on 242 degrees of freedom
Residual deviance: 4.638 on 214 degrees of freedom
AIC: 1855.6
Number of Fisher Scoring iterations: 4Seleção das variáveis
#step(mlg3)
mlg4.1 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m + Carbono,
family = Gamma(link = "inverse"), data = dados[, c(-2, -3, -30)])
summary(mlg4.1)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m + Carbono,
family = Gamma(link = "inverse"), data = dados[, c(-2, -3,
-30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.61877 -0.07334 0.00792 0.07941 0.40846
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.049e-02 3.114e-04 33.706 < 2e-16 ***
La 2.688e-02 1.439e-02 1.867 0.06319 .
Ce 2.691e-02 1.439e-02 1.869 0.06290 .
Pr 2.699e-02 1.439e-02 1.876 0.06198 .
Nd 2.694e-02 1.439e-02 1.872 0.06255 .
Sm 2.709e-02 1.445e-02 1.874 0.06221 .
Eu 2.793e-02 1.434e-02 1.948 0.05271 .
Gd 2.630e-02 1.435e-02 1.832 0.06822 .
Tb 2.369e-02 1.420e-02 1.668 0.09675 .
Dy 2.677e-02 1.439e-02 1.860 0.06419 .
Er 2.916e-02 1.468e-02 1.987 0.04811 *
Yb 2.504e-02 1.426e-02 1.756 0.08047 .
Lu 3.216e-02 1.447e-02 2.222 0.02731 *
ETRs -2.690e-02 1.439e-02 -1.869 0.06295 .
Ca 3.587e-04 1.436e-04 2.498 0.01323 *
Mg 4.581e-04 1.945e-04 2.356 0.01936 *
P -4.021e-05 1.385e-05 -2.904 0.00405 **
Al 1.706e-03 3.412e-04 5.001 1.15e-06 ***
m -3.049e-04 1.300e-04 -2.346 0.01984 *
Carbono 4.437e-04 2.120e-04 2.093 0.03746 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.01935132)
Null deviance: 13.4192 on 242 degrees of freedom
Residual deviance: 4.7246 on 223 degrees of freedom
AIC: 1842.1
Number of Fisher Scoring iterations: 4mlg4.2 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd +
Dy + Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m + Carbono,
family = Gamma(link = "inverse"), data = dados[, c(-2, -3, -30)])
summary(mlg4.2)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Dy +
Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m + Carbono, family = Gamma(link = "inverse"),
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.60756 -0.06357 0.01040 0.07491 0.41332
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.070e-02 2.886e-04 37.069 < 2e-16 ***
La 3.033e-03 1.659e-03 1.828 0.06883 .
Ce 3.062e-03 1.661e-03 1.843 0.06660 .
Pr 3.162e-03 1.689e-03 1.872 0.06249 .
Nd 3.097e-03 1.655e-03 1.872 0.06252 .
Sm 3.178e-03 1.822e-03 1.744 0.08254 .
Eu 4.178e-03 1.673e-03 2.497 0.01326 *
Gd 2.533e-03 1.693e-03 1.496 0.13601
Dy 2.935e-03 1.681e-03 1.746 0.08210 .
Er 5.034e-03 2.457e-03 2.049 0.04163 *
Yb 1.390e-03 1.506e-03 0.923 0.35708
Lu 8.570e-03 3.068e-03 2.794 0.00566 **
ETRs -3.057e-03 1.661e-03 -1.840 0.06705 .
Ca 3.775e-04 1.437e-04 2.626 0.00922 **
Mg 4.157e-04 1.924e-04 2.161 0.03176 *
P -3.722e-05 1.367e-05 -2.722 0.00700 **
Al 1.784e-03 3.411e-04 5.232 3.85e-07 ***
m -3.575e-04 1.268e-04 -2.818 0.00526 **
Carbono 4.994e-04 2.111e-04 2.365 0.01886 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.0195601)
Null deviance: 13.4192 on 242 degrees of freedom
Residual deviance: 4.7789 on 224 degrees of freedom
AIC: 1842.9
Number of Fisher Scoring iterations: 4mlg4.3 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd +
Dy + Er + Lu + ETRs + Ca + Mg + P + Al + m + Carbono,
family = Gamma(link = "inverse"), data = dados[, c(-2, -3, -30)])
summary(mlg4.3)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Dy +
Er + Lu + ETRs + Ca + Mg + P + Al + m + Carbono, family = Gamma(link = "inverse"),
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.60433 -0.06254 0.01435 0.07275 0.41408
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.072e-02 2.882e-04 37.190 < 2e-16 ***
La 1.920e-03 1.133e-03 1.694 0.09158 .
Ce 1.950e-03 1.137e-03 1.715 0.08780 .
Pr 2.027e-03 1.152e-03 1.760 0.07975 .
Nd 1.996e-03 1.140e-03 1.751 0.08136 .
Sm 1.921e-03 1.203e-03 1.596 0.11180
Eu 3.181e-03 1.276e-03 2.493 0.01339 *
Gd 1.410e-03 1.172e-03 1.203 0.23009
Dy 1.813e-03 1.154e-03 1.572 0.11747
Er 4.089e-03 2.224e-03 1.838 0.06733 .
Lu 6.847e-03 2.416e-03 2.834 0.00501 **
ETRs -1.944e-03 1.136e-03 -1.711 0.08849 .
Ca 3.840e-04 1.442e-04 2.664 0.00829 **
Mg 4.040e-04 1.930e-04 2.093 0.03743 *
P -3.645e-05 1.372e-05 -2.656 0.00847 **
Al 1.795e-03 3.412e-04 5.262 3.32e-07 ***
m -3.456e-04 1.264e-04 -2.734 0.00675 **
Carbono 5.038e-04 2.116e-04 2.381 0.01809 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.01962428)
Null deviance: 13.4192 on 242 degrees of freedom
Residual deviance: 4.7956 on 225 degrees of freedom
AIC: 1841.8
Number of Fisher Scoring iterations: 4mlg4.4 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu +
Dy + Er + Lu + ETRs + Ca + Mg + P + Al + m + Carbono,
family = Gamma(link = "inverse"), data = dados[, c(-2, -3, -30)])
summary(mlg4.4)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Dy + Er +
Lu + ETRs + Ca + Mg + P + Al + m + Carbono, family = Gamma(link = "inverse"),
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.60522 -0.06429 0.01635 0.07609 0.43143
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.074e-02 2.875e-04 37.357 < 2e-16 ***
La 5.683e-04 1.505e-04 3.776 0.000204 ***
Ce 5.929e-04 1.487e-04 3.987 9.02e-05 ***
Pr 6.553e-04 1.666e-04 3.934 0.000111 ***
Nd 6.403e-04 1.773e-04 3.611 0.000376 ***
Sm 5.599e-04 4.113e-04 1.361 0.174788
Eu 1.811e-03 5.647e-04 3.207 0.001538 **
Dy 4.402e-04 1.781e-04 2.472 0.014172 *
Er 1.536e-03 6.639e-04 2.313 0.021620 *
Lu 4.128e-03 8.371e-04 4.932 1.58e-06 ***
ETRs -5.882e-04 1.498e-04 -3.927 0.000114 ***
Ca 3.985e-04 1.439e-04 2.770 0.006072 **
Mg 4.034e-04 1.940e-04 2.080 0.038669 *
P -3.677e-05 1.379e-05 -2.666 0.008239 **
Al 1.862e-03 3.361e-04 5.539 8.41e-08 ***
m -3.268e-04 1.253e-04 -2.609 0.009678 **
Carbono 4.670e-04 2.093e-04 2.231 0.026638 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.01962408)
Null deviance: 13.419 on 242 degrees of freedom
Residual deviance: 4.824 on 226 degrees of freedom
AIC: 1841.2
Number of Fisher Scoring iterations: 4mlg4.5 = glm(formula = Areia ~ La + Ce + Pr + Nd + Eu +
Dy + Er + Lu + ETRs + Ca + Mg + P + Al + m + Carbono,
family = Gamma(link = "inverse"), data = dados[, c(-2, -3, -30)])
summary(mlg4.5)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Eu + Dy + Er + Lu +
ETRs + Ca + Mg + P + Al + m + Carbono, family = Gamma(link = "inverse"),
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.60076 -0.06295 0.01418 0.07737 0.43087
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.081e-02 2.825e-04 38.268 < 2e-16 ***
La 4.207e-04 1.030e-04 4.085 6.11e-05 ***
Ce 4.642e-04 1.135e-04 4.089 6.02e-05 ***
Pr 5.643e-04 1.511e-04 3.734 0.000238 ***
Nd 5.736e-04 1.706e-04 3.362 0.000908 ***
Eu 1.766e-03 5.601e-04 3.154 0.001830 **
Dy 3.326e-04 1.600e-04 2.078 0.038788 *
Er 1.387e-03 6.530e-04 2.124 0.034771 *
Lu 4.018e-03 8.329e-04 4.823 2.59e-06 ***
ETRs -4.594e-04 1.150e-04 -3.997 8.69e-05 ***
Ca 3.924e-04 1.441e-04 2.723 0.006971 **
Mg 4.081e-04 1.940e-04 2.104 0.036519 *
P -3.624e-05 1.379e-05 -2.627 0.009196 **
Al 1.876e-03 3.354e-04 5.593 6.39e-08 ***
m -3.309e-04 1.250e-04 -2.648 0.008661 **
Carbono 4.548e-04 2.092e-04 2.173 0.030781 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.01962367)
Null deviance: 13.4192 on 242 degrees of freedom
Residual deviance: 4.8604 on 227 degrees of freedom
AIC: 1841
Number of Fisher Scoring iterations: 4Verificando os resíduos do mlg4.5
par(mfrow = c(2,2))
plot(mlg4.5, pch = 20)
Envelope
set.seed(1234)
hnp(mlg4.5, print.on = T, main="Areia")Gamma model 
Distribuição Normal inversa para a Variável Areia
mlg5 = glm(Areia~., data = dados[,c(-2,-3,-30)], family = inverse.gaussian)
summary(mlg5)
Call:
glm(formula = Areia ~ ., family = inverse.gaussian, data = dados[,
c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.092372 -0.008363 0.001161 0.009739 0.049606
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.365e-04 1.160e-04 1.177 0.240628
La 5.564e-04 7.192e-04 0.774 0.439973
Ce 5.571e-04 7.193e-04 0.775 0.439465
Pr 5.592e-04 7.194e-04 0.777 0.437839
Nd 5.592e-04 7.194e-04 0.777 0.437888
Sm 5.582e-04 7.197e-04 0.776 0.438788
Eu 5.940e-04 7.181e-04 0.827 0.409050
Gd 5.536e-04 1.058e-03 0.523 0.601366
Tb 5.269e-04 1.061e-03 0.497 0.620000
Dy 5.728e-04 1.058e-03 0.541 0.588933
Er 6.380e-04 1.058e-03 0.603 0.547100
Yb 5.373e-04 1.062e-03 0.506 0.613485
Lu 7.091e-04 1.056e-03 0.672 0.502428
ETRLs 9.202e-04 1.542e-03 0.597 0.551367
ETRPs 8.999e-04 1.695e-03 0.531 0.596092
ETRs -1.477e-03 1.338e-03 -1.104 0.270665
ETRLs_ETRPs -5.874e-08 5.227e-07 -0.112 0.910628
Ca 1.449e-05 5.557e-06 2.607 0.009769 **
Mg 1.213e-05 6.186e-06 1.962 0.051116 .
P -1.144e-06 4.002e-07 -2.858 0.004689 **
K 1.385e-07 1.957e-07 0.708 0.479927
Al 5.247e-05 1.378e-05 3.809 0.000183 ***
H_Al -7.398e-04 8.211e-04 -0.901 0.368614
SB -7.404e-04 8.208e-04 -0.902 0.368053
V -4.248e-07 1.103e-06 -0.385 0.700554
T 7.403e-04 8.208e-04 0.902 0.368136
m -1.154e-05 6.538e-06 -1.765 0.078908 .
Carbono 8.230e-06 6.964e-06 1.182 0.238637
pH 1.387e-06 6.676e-06 0.208 0.835568
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for inverse.gaussian family taken to be 0.0003550614)
Null deviance: 0.226095 on 242 degrees of freedom
Residual deviance: 0.089344 on 214 degrees of freedom
AIC: 1928.7
Number of Fisher Scoring iterations: 5Seleção das Variáveis
#step(mlg5)
mlg5.1 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
summary(mlg5.1)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Tb +
Dy + Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.091640 -0.007083 0.001729 0.010165 0.048434
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.058e-04 8.838e-06 11.971 < 2e-16 ***
La 8.190e-04 4.496e-04 1.822 0.06983 .
Ce 8.198e-04 4.496e-04 1.824 0.06956 .
Pr 8.222e-04 4.494e-04 1.830 0.06863 .
Nd 8.222e-04 4.496e-04 1.829 0.06878 .
Sm 8.200e-04 4.515e-04 1.816 0.07068 .
Eu 8.595e-04 4.477e-04 1.920 0.05617 .
Gd 7.954e-04 4.482e-04 1.775 0.07729 .
Tb 7.607e-04 4.427e-04 1.718 0.08712 .
Dy 8.153e-04 4.495e-04 1.814 0.07107 .
Er 8.833e-04 4.605e-04 1.918 0.05637 .
Yb 7.783e-04 4.441e-04 1.752 0.08107 .
Lu 9.608e-04 4.522e-04 2.125 0.03470 *
ETRs -8.198e-04 4.496e-04 -1.824 0.06955 .
Ca 1.507e-05 4.802e-06 3.138 0.00193 **
Mg 1.154e-05 5.524e-06 2.088 0.03789 *
P -1.138e-06 3.552e-07 -3.205 0.00155 **
Al 4.819e-05 1.099e-05 4.387 1.77e-05 ***
m -7.113e-06 4.029e-06 -1.765 0.07885 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for inverse.gaussian family taken to be 0.000345206)
Null deviance: 0.226095 on 242 degrees of freedom
Residual deviance: 0.090943 on 224 degrees of freedom
AIC: 1913
Number of Fisher Scoring iterations: 5mlg5.2 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd +
Dy + Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
summary(mlg5.2)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Dy +
Er + Yb + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.093236 -0.007021 0.002262 0.009971 0.049593
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.131e-04 7.875e-06 14.357 < 2e-16 ***
La 5.272e-05 5.589e-05 0.943 0.346561
Ce 5.354e-05 5.598e-05 0.956 0.339939
Pr 5.650e-05 5.694e-05 0.992 0.322104
Nd 5.585e-05 5.585e-05 1.000 0.318391
Sm 5.148e-05 6.092e-05 0.845 0.398942
Eu 9.644e-05 5.666e-05 1.702 0.090147 .
Gd 3.196e-05 5.755e-05 0.555 0.579225
Dy 4.934e-05 5.675e-05 0.869 0.385585
Er 1.046e-04 8.181e-05 1.278 0.202435
Yb 2.073e-05 5.312e-05 0.390 0.696730
Lu 2.052e-04 1.042e-04 1.970 0.050044 .
ETRs -5.355e-05 5.598e-05 -0.957 0.339781
Ca 1.616e-05 4.789e-06 3.374 0.000873 ***
Mg 9.708e-06 5.417e-06 1.792 0.074486 .
P -1.021e-06 3.471e-07 -2.943 0.003594 **
Al 5.086e-05 1.097e-05 4.636 6.02e-06 ***
m -8.593e-06 3.953e-06 -2.174 0.030774 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for inverse.gaussian family taken to be 0.0003482664)
Null deviance: 0.22609 on 242 degrees of freedom
Residual deviance: 0.09198 on 225 degrees of freedom
AIC: 1913.8
Number of Fisher Scoring iterations: 5mlg5.3 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd +
Dy + Er + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
summary(mlg4.3)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Gd + Dy +
Er + Lu + ETRs + Ca + Mg + P + Al + m + Carbono, family = Gamma(link = "inverse"),
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.60433 -0.06254 0.01435 0.07275 0.41408
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.072e-02 2.882e-04 37.190 < 2e-16 ***
La 1.920e-03 1.133e-03 1.694 0.09158 .
Ce 1.950e-03 1.137e-03 1.715 0.08780 .
Pr 2.027e-03 1.152e-03 1.760 0.07975 .
Nd 1.996e-03 1.140e-03 1.751 0.08136 .
Sm 1.921e-03 1.203e-03 1.596 0.11180
Eu 3.181e-03 1.276e-03 2.493 0.01339 *
Gd 1.410e-03 1.172e-03 1.203 0.23009
Dy 1.813e-03 1.154e-03 1.572 0.11747
Er 4.089e-03 2.224e-03 1.838 0.06733 .
Lu 6.847e-03 2.416e-03 2.834 0.00501 **
ETRs -1.944e-03 1.136e-03 -1.711 0.08849 .
Ca 3.840e-04 1.442e-04 2.664 0.00829 **
Mg 4.040e-04 1.930e-04 2.093 0.03743 *
P -3.645e-05 1.372e-05 -2.656 0.00847 **
Al 1.795e-03 3.412e-04 5.262 3.32e-07 ***
m -3.456e-04 1.264e-04 -2.734 0.00675 **
Carbono 5.038e-04 2.116e-04 2.381 0.01809 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.01962428)
Null deviance: 13.4192 on 242 degrees of freedom
Residual deviance: 4.7956 on 225 degrees of freedom
AIC: 1841.8
Number of Fisher Scoring iterations: 4mlg5.4 = glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu +
Dy + Er + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
summary(mlg5.4)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Sm + Eu + Dy + Er +
Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.092186 -0.007295 0.002029 0.009853 0.051018
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.134e-04 7.818e-06 14.504 < 2e-16 ***
La 2.185e-05 5.366e-06 4.071 6.47e-05 ***
Ce 2.261e-05 5.290e-06 4.273 2.84e-05 ***
Pr 2.507e-05 5.853e-06 4.283 2.72e-05 ***
Nd 2.504e-05 6.228e-06 4.021 7.90e-05 ***
Sm 1.877e-05 1.443e-05 1.301 0.194597
Eu 6.684e-05 1.901e-05 3.515 0.000531 ***
Dy 1.803e-05 6.325e-06 2.851 0.004759 **
Er 6.388e-05 2.326e-05 2.746 0.006508 **
Lu 1.498e-04 2.678e-05 5.594 6.36e-08 ***
ETRs -2.263e-05 5.335e-06 -4.241 3.23e-05 ***
Ca 1.635e-05 4.771e-06 3.428 0.000723 ***
Mg 9.636e-06 5.410e-06 1.781 0.076217 .
P -1.021e-06 3.467e-07 -2.944 0.003572 **
Al 5.178e-05 1.077e-05 4.809 2.76e-06 ***
m -8.322e-06 3.914e-06 -2.126 0.034583 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for inverse.gaussian family taken to be 0.0003468838)
Null deviance: 0.226095 on 242 degrees of freedom
Residual deviance: 0.092088 on 227 degrees of freedom
AIC: 1910.1
Number of Fisher Scoring iterations: 5mlg5.5 = glm(formula = Areia ~ La + Ce + Pr + Nd + Eu +
Dy + Er + Lu + ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
summary(mlg5.5)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Eu + Dy + Er + Lu +
ETRs + Ca + Mg + P + Al + m, family = inverse.gaussian, data = dados[,
c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.092729 -0.006876 0.001594 0.009709 0.050154
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.153e-04 7.646e-06 15.085 < 2e-16 ***
La 1.676e-05 3.612e-06 4.640 5.88e-06 ***
Ce 1.814e-05 3.966e-06 4.575 7.82e-06 ***
Pr 2.179e-05 5.214e-06 4.180 4.16e-05 ***
Nd 2.269e-05 5.951e-06 3.812 0.000177 ***
Eu 6.477e-05 1.880e-05 3.445 0.000679 ***
Dy 1.455e-05 5.757e-06 2.528 0.012149 *
Er 5.764e-05 2.265e-05 2.544 0.011605 *
Lu 1.448e-04 2.644e-05 5.475 1.15e-07 ***
ETRs -1.816e-05 4.028e-06 -4.509 1.04e-05 ***
Ca 1.621e-05 4.771e-06 3.399 0.000799 ***
Mg 9.852e-06 5.405e-06 1.823 0.069670 .
P -1.014e-06 3.460e-07 -2.932 0.003713 **
Al 5.223e-05 1.072e-05 4.871 2.08e-06 ***
m -8.429e-06 3.893e-06 -2.165 0.031407 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for inverse.gaussian family taken to be 0.0003457877)
Null deviance: 0.226095 on 242 degrees of freedom
Residual deviance: 0.092675 on 228 degrees of freedom
AIC: 1909.6
Number of Fisher Scoring iterations: 5mlg5.6 = glm(formula = Areia ~ La + Ce + Pr + Nd + Eu +
Dy + Er + Lu + ETRs + Ca + P + Al + m, family = inverse.gaussian,
data = dados[, c(-2, -3, -30)])
summary(mlg5.6)
Call:
glm(formula = Areia ~ La + Ce + Pr + Nd + Eu + Dy + Er + Lu +
ETRs + Ca + P + Al + m, family = inverse.gaussian, data = dados[,
c(-2, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-0.100814 -0.007189 0.001447 0.009551 0.049703
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.110e-04 7.202e-06 15.410 < 2e-16 ***
La 1.738e-05 3.585e-06 4.847 2.31e-06 ***
Ce 1.874e-05 3.939e-06 4.757 3.48e-06 ***
Pr 2.208e-05 5.208e-06 4.240 3.24e-05 ***
Nd 2.325e-05 5.935e-06 3.917 0.000118 ***
Eu 6.884e-05 1.871e-05 3.680 0.000291 ***
Dy 1.453e-05 5.768e-06 2.518 0.012481 *
Er 6.219e-05 2.248e-05 2.766 0.006134 **
Lu 1.479e-04 2.644e-05 5.596 6.25e-08 ***
ETRs -1.872e-05 4.006e-06 -4.674 5.04e-06 ***
Ca 2.191e-05 3.584e-06 6.112 4.18e-09 ***
P -4.842e-07 2.344e-07 -2.066 0.039945 *
Al 5.509e-05 1.063e-05 5.182 4.81e-07 ***
m -8.791e-06 3.901e-06 -2.253 0.025193 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for inverse.gaussian family taken to be 0.0003461534)
Null deviance: 0.226095 on 242 degrees of freedom
Residual deviance: 0.093724 on 229 degrees of freedom
AIC: 1910.4
Number of Fisher Scoring iterations: 5Verificando os resíduos do mlg5.6
par(mfrow = c(2,2))
plot(mlg5.6, pch = 20)
Envelope
set.seed(1435)
hnp(mlg5.6, print= T, main="Areia")Inverse gaussian model 
Escolhendo melhor modelo para Var. Areia
ajuste = c('mlg4.5','mlg5.6')
aic = c(AIC(mlg4.5), AIC(mlg5.6))
deviance = c(deviance(mlg4.5),deviance(mlg5.6))
verossimilhanca =c(logLik(mlg4.5),logLik(mlg5.6))
data.frame(ajuste, aic, verossimilhanca,deviance) ajuste aic verossimilhanca deviance
1 mlg4.5 1841.041 -903.5205 4.86039336
2 mlg5.6 1910.354 -940.1772 0.09372384Distribuição Gamma para a variável Argila
mlg6 = glm(Argila~., data = dados[,c(-1,-3,-30)], family = Gamma(link="inverse"))
summary(mlg6)
Call:
glm(formula = Argila ~ ., family = Gamma(link = "inverse"), data = dados[,
c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.62662 -0.30765 -0.05476 0.22112 1.68064
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.072e-02 5.381e-02 0.199 0.842348
La 4.485e-01 3.280e-01 1.367 0.172966
Ce 4.485e-01 3.280e-01 1.367 0.173029
Pr 4.460e-01 3.281e-01 1.359 0.175492
Nd 4.494e-01 3.281e-01 1.370 0.172219
Sm 4.412e-01 3.282e-01 1.344 0.180244
Eu 4.461e-01 3.276e-01 1.362 0.174710
Gd -1.810e-02 4.494e-01 -0.040 0.967904
Tb 2.724e-02 4.495e-01 0.061 0.951739
Dy -2.036e-02 4.496e-01 -0.045 0.963928
Er -2.085e-02 4.496e-01 -0.046 0.963050
Yb -2.034e-02 4.498e-01 -0.045 0.963970
Lu -4.889e-02 4.492e-01 -0.109 0.913427
ETRLs -9.728e-01 6.875e-01 -1.415 0.158558
ETRPs -5.047e-01 6.883e-01 -0.733 0.464200
ETRs 5.243e-01 5.483e-01 0.956 0.340028
ETRLs_ETRPs 6.463e-04 2.837e-04 2.278 0.023707 *
Ca 4.804e-04 8.996e-04 0.534 0.593843
Mg -1.758e-03 1.298e-03 -1.354 0.177139
P 1.341e-04 1.100e-04 1.219 0.224248
K -1.556e-04 6.901e-05 -2.255 0.025117 *
Al -7.514e-03 3.863e-03 -1.945 0.053055 .
H_Al -6.490e-02 3.724e-01 -0.174 0.861815
SB -6.371e-02 3.721e-01 -0.171 0.864213
V 7.254e-06 5.014e-04 0.014 0.988470
T 6.381e-02 3.721e-01 0.171 0.864010
m 2.612e-03 2.745e-03 0.952 0.342381
Carbono -5.595e-03 3.076e-03 -1.819 0.070326 .
pH 1.316e-02 3.762e-03 3.500 0.000566 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2346213)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 66.055 on 214 degrees of freedom
AIC: 1733.2
Number of Fisher Scoring iterations: 5Seleção das variáveis
#step(mlg6)
mlg6.1 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + H_Al + T + m + Carbono +
pH, family = Gamma(link = "inverse"), data = dados[, c(-1,-3, -30)])
summary(mlg6.1)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + H_Al + T + m + Carbono + pH,
family = Gamma(link = "inverse"), data = dados[, c(-1, -3,
-30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.64598 -0.31509 -0.05439 0.21422 1.61380
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.562e-02 1.513e-02 1.032 0.303118
La 5.737e-01 2.801e-01 2.048 0.041707 *
Ce 5.737e-01 2.801e-01 2.048 0.041690 *
Pr 5.719e-01 2.801e-01 2.042 0.042334 *
Nd 5.749e-01 2.801e-01 2.052 0.041283 *
Sm 5.689e-01 2.801e-01 2.031 0.043397 *
Eu 5.649e-01 2.806e-01 2.013 0.045267 *
Tb 3.100e-02 1.298e-02 2.389 0.017725 *
Lu -3.200e-02 1.103e-02 -2.901 0.004087 **
ETRLs -5.738e-01 2.801e-01 -2.048 0.041678 *
ETRLs_ETRPs 4.947e-04 2.539e-04 1.948 0.052605 .
K -1.590e-04 6.224e-05 -2.555 0.011265 *
Al -8.464e-03 2.945e-03 -2.874 0.004447 **
H_Al -8.199e-04 5.099e-04 -1.608 0.109204
T 9.355e-05 5.959e-05 1.570 0.117864
m 2.778e-03 1.476e-03 1.882 0.061114 .
Carbono -5.053e-03 2.634e-03 -1.918 0.056329 .
pH 1.228e-02 3.191e-03 3.848 0.000155 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2168136)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 67.434 on 225 degrees of freedom
AIC: 1716.4
Number of Fisher Scoring iterations: 5mlg6.2 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + H_Al + m + Carbono +
pH, family = Gamma(link = "inverse"), data = dados[, c(-1,-3, -30)])
summary(mlg6.2)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + H_Al + m + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.63626 -0.31097 -0.02913 0.21558 1.63288
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.288e-02 1.496e-02 0.861 0.39032
La 5.345e-01 2.814e-01 1.899 0.05879 .
Ce 5.344e-01 2.814e-01 1.899 0.05884 .
Pr 5.325e-01 2.814e-01 1.892 0.05973 .
Nd 5.351e-01 2.814e-01 1.902 0.05848 .
Sm 5.320e-01 2.815e-01 1.890 0.06004 .
Eu 5.234e-01 2.818e-01 1.858 0.06453 .
Tb 2.744e-02 1.261e-02 2.177 0.03055 *
Lu -3.055e-02 1.110e-02 -2.753 0.00639 **
ETRLs -5.344e-01 2.814e-01 -1.899 0.05882 .
ETRLs_ETRPs 4.422e-04 2.549e-04 1.734 0.08420 .
K -7.784e-05 3.471e-05 -2.243 0.02587 *
Al -7.615e-03 2.797e-03 -2.723 0.00697 **
H_Al -6.488e-04 5.084e-04 -1.276 0.20325
m 2.149e-03 1.424e-03 1.509 0.13276
Carbono -4.379e-03 2.628e-03 -1.667 0.09699 .
pH 1.318e-02 3.119e-03 4.226 3.45e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2190851)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 68.012 on 226 degrees of freedom
AIC: 1716.6
Number of Fisher Scoring iterations: 5mlg6.3 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + m + Carbono +
pH, family = Gamma(link = "inverse"), data = dados[, c(-1,-3, -30)])
summary(mlg6.3)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + m + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.63744 -0.32313 -0.04166 0.21446 1.59019
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.223e-02 1.506e-02 0.812 0.417491
La 5.262e-01 2.802e-01 1.878 0.061726 .
Ce 5.260e-01 2.802e-01 1.877 0.061774 .
Pr 5.241e-01 2.802e-01 1.870 0.062717 .
Nd 5.269e-01 2.802e-01 1.880 0.061350 .
Sm 5.225e-01 2.803e-01 1.865 0.063539 .
Eu 5.154e-01 2.806e-01 1.837 0.067544 .
Tb 3.190e-02 1.203e-02 2.652 0.008569 **
Lu -3.531e-02 1.043e-02 -3.386 0.000836 ***
ETRLs -5.261e-01 2.802e-01 -1.877 0.061759 .
ETRLs_ETRPs 3.708e-04 2.507e-04 1.479 0.140531
K -7.721e-05 3.525e-05 -2.191 0.029504 *
Al -8.447e-03 2.718e-03 -3.108 0.002127 **
m 2.041e-03 1.446e-03 1.412 0.159397
Carbono -4.333e-03 2.625e-03 -1.651 0.100161
pH 1.340e-02 3.144e-03 4.262 2.98e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2196031)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 68.341 on 227 degrees of freedom
AIC: 1715.8
Number of Fisher Scoring iterations: 5step(mlg6.3)Start: AIC=1715.8
Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu + ETRLs + ETRLs_ETRPs +
K + Al + m + Carbono + pH
Df Deviance AIC
<none> 68.341 1715.8
- m 1 68.788 1715.8
- ETRLs_ETRPs 1 68.852 1716.1
- Carbono 1 68.923 1716.5
- Eu 1 69.097 1717.2
- Sm 1 69.120 1717.3
- Pr 1 69.125 1717.4
- Ce 1 69.131 1717.4
- ETRLs 1 69.131 1717.4
- La 1 69.131 1717.4
- Nd 1 69.134 1717.4
- K 1 69.294 1718.1
- Tb 1 69.915 1721.0
- Al 1 69.925 1721.0
- Lu 1 70.832 1725.2
- pH 1 72.540 1732.9
Call: glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + m + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Coefficients:
(Intercept) La Ce Pr Nd Sm
1.223e-02 5.262e-01 5.260e-01 5.241e-01 5.269e-01 5.225e-01
Eu Tb Lu ETRLs ETRLs_ETRPs K
5.154e-01 3.190e-02 -3.531e-02 -5.261e-01 3.708e-04 -7.721e-05
Al m Carbono pH
-8.447e-03 2.041e-03 -4.333e-03 1.340e-02
Degrees of Freedom: 242 Total (i.e. Null); 227 Residual
Null Deviance: 94.64
Residual Deviance: 68.34 AIC: 1716mlg6.4 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + m + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.4)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + m + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.63744 -0.32313 -0.04166 0.21446 1.59019
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.223e-02 1.506e-02 0.812 0.417491
La 5.262e-01 2.802e-01 1.878 0.061726 .
Ce 5.260e-01 2.802e-01 1.877 0.061774 .
Pr 5.241e-01 2.802e-01 1.870 0.062717 .
Nd 5.269e-01 2.802e-01 1.880 0.061350 .
Sm 5.225e-01 2.803e-01 1.865 0.063539 .
Eu 5.154e-01 2.806e-01 1.837 0.067544 .
Tb 3.190e-02 1.203e-02 2.652 0.008569 **
Lu -3.531e-02 1.043e-02 -3.386 0.000836 ***
ETRLs -5.261e-01 2.802e-01 -1.877 0.061759 .
ETRLs_ETRPs 3.708e-04 2.507e-04 1.479 0.140531
K -7.721e-05 3.525e-05 -2.191 0.029504 *
Al -8.447e-03 2.718e-03 -3.108 0.002127 **
m 2.041e-03 1.446e-03 1.412 0.159397
Carbono -4.333e-03 2.625e-03 -1.651 0.100161
pH 1.340e-02 3.144e-03 4.262 2.98e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2196031)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 68.341 on 227 degrees of freedom
AIC: 1715.8
Number of Fisher Scoring iterations: 5mlg6.5 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.5)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + Carbono + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.63712 -0.32798 -0.02304 0.20939 1.58982
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.589e-02 1.503e-02 1.057 0.29162
La 4.175e-01 2.714e-01 1.538 0.12533
Ce 4.174e-01 2.714e-01 1.538 0.12542
Pr 4.158e-01 2.714e-01 1.532 0.12692
Nd 4.180e-01 2.713e-01 1.541 0.12479
Sm 4.151e-01 2.716e-01 1.528 0.12786
Eu 4.063e-01 2.716e-01 1.496 0.13603
Tb 2.318e-02 1.040e-02 2.230 0.02675 *
Lu -2.657e-02 8.373e-03 -3.174 0.00171 **
ETRLs -4.174e-01 2.714e-01 -1.538 0.12540
ETRLs_ETRPs 3.551e-04 2.537e-04 1.399 0.16303
K -9.685e-05 3.214e-05 -3.013 0.00288 **
Al -6.731e-03 2.712e-03 -2.482 0.01379 *
Carbono -3.631e-03 2.608e-03 -1.392 0.16527
pH 1.302e-02 3.164e-03 4.113 5.44e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2214802)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 68.788 on 228 degrees of freedom
AIC: 1715.5
Number of Fisher Scoring iterations: 5mlg6.6 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.6)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Eu + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.6380 -0.3343 -0.0286 0.2173 1.5367
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 8.373e-03 1.414e-02 0.592 0.554309
La 3.981e-01 2.716e-01 1.466 0.144073
Ce 3.980e-01 2.716e-01 1.465 0.144192
Pr 3.963e-01 2.716e-01 1.459 0.145855
Nd 3.985e-01 2.715e-01 1.468 0.143533
Sm 3.956e-01 2.718e-01 1.455 0.147012
Eu 3.870e-01 2.718e-01 1.424 0.155911
Tb 2.564e-02 1.023e-02 2.506 0.012921 *
Lu -2.990e-02 8.027e-03 -3.725 0.000246 ***
ETRLs -3.980e-01 2.716e-01 -1.465 0.144161
ETRLs_ETRPs 4.118e-04 2.575e-04 1.599 0.111119
K -9.179e-05 3.256e-05 -2.819 0.005237 **
Al -6.534e-03 2.709e-03 -2.412 0.016645 *
pH 1.379e-02 3.157e-03 4.368 1.9e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2251895)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 69.208 on 229 degrees of freedom
AIC: 1715
Number of Fisher Scoring iterations: 5mlg6.7 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Tb + Lu +
ETRLs + ETRLs_ETRPs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.7)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Tb + Lu + ETRLs +
ETRLs_ETRPs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.64397 -0.32690 -0.02512 0.20214 1.58587
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.853e-02 1.229e-02 1.507 0.13307
La 1.158e-02 5.726e-03 2.022 0.04436 *
Ce 1.146e-02 5.875e-03 1.951 0.05224 .
Pr 9.807e-03 5.645e-03 1.737 0.08368 .
Nd 1.214e-02 6.268e-03 1.936 0.05403 .
Sm 8.669e-03 6.516e-03 1.330 0.18468
Tb 2.696e-02 1.043e-02 2.585 0.01035 *
Lu -3.022e-02 7.967e-03 -3.793 0.00019 ***
ETRLs -1.149e-02 5.864e-03 -1.960 0.05119 .
ETRLs_ETRPs 3.756e-04 2.565e-04 1.464 0.14451
K -7.929e-05 3.264e-05 -2.429 0.01591 *
Al -7.149e-03 2.656e-03 -2.692 0.00763 **
pH 1.221e-02 2.948e-03 4.143 4.82e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2274487)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 69.673 on 230 degrees of freedom
AIC: 1714.7
Number of Fisher Scoring iterations: 5mlg6.8 = glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Tb + Lu +
ETRLs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.8)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Sm + Tb + Lu + ETRLs +
K + Al + pH, family = Gamma(link = "inverse"), data = dados[,
c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.63479 -0.34266 -0.03149 0.21675 1.57630
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.009e-02 1.243e-02 1.616 0.10755
La 1.173e-02 5.564e-03 2.108 0.03615 *
Ce 1.160e-02 5.717e-03 2.030 0.04354 *
Pr 1.012e-02 5.497e-03 1.840 0.06700 .
Nd 1.217e-02 6.109e-03 1.993 0.04743 *
Sm 8.669e-03 6.329e-03 1.370 0.17211
Tb 2.632e-02 1.032e-02 2.551 0.01138 *
Lu -3.345e-02 7.681e-03 -4.355 2.00e-05 ***
ETRLs -1.162e-02 5.706e-03 -2.036 0.04291 *
K -8.836e-05 3.181e-05 -2.778 0.00592 **
Al -6.547e-03 2.710e-03 -2.416 0.01647 *
pH 1.347e-02 2.860e-03 4.710 4.28e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2264637)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 70.187 on 231 degrees of freedom
AIC: 1714.6
Number of Fisher Scoring iterations: 5mlg6.9 = glm(formula = Argila ~ La + Ce + Pr + Nd + Tb + Lu +
ETRLs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.9)
Call:
glm(formula = Argila ~ La + Ce + Pr + Nd + Tb + Lu + ETRLs +
K + Al + pH, family = Gamma(link = "inverse"), data = dados[,
c(-1, -3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.64587 -0.32766 -0.03642 0.22688 1.55746
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.433e-02 1.208e-02 2.013 0.04522 *
La 4.935e-03 2.489e-03 1.983 0.04857 *
Ce 4.693e-03 2.706e-03 1.734 0.08426 .
Pr 3.572e-03 2.723e-03 1.312 0.19089
Nd 5.189e-03 3.431e-03 1.512 0.13179
Tb 2.849e-02 9.948e-03 2.864 0.00457 **
Lu -3.693e-02 7.134e-03 -5.177 4.89e-07 ***
ETRLs -4.724e-03 2.700e-03 -1.750 0.08149 .
K -8.404e-05 3.147e-05 -2.671 0.00811 **
Al -6.374e-03 2.764e-03 -2.306 0.02197 *
pH 1.246e-02 2.756e-03 4.519 9.91e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2265467)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 70.649 on 232 degrees of freedom
AIC: 1714.3
Number of Fisher Scoring iterations: 5mlg6.10 = glm(formula = Argila ~ La + Ce + Nd + Tb + Lu +
ETRLs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.10)
Call:
glm(formula = Argila ~ La + Ce + Nd + Tb + Lu + ETRLs + K + Al +
pH, family = Gamma(link = "inverse"), data = dados[, c(-1,
-3, -30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.65594 -0.33585 -0.04167 0.23397 1.52610
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.879e-02 1.162e-02 2.478 0.01391 *
La 1.824e-03 7.333e-04 2.488 0.01355 *
Ce 1.253e-03 6.477e-04 1.935 0.05425 .
Nd 8.783e-04 9.401e-04 0.934 0.35114
Tb 1.945e-02 6.926e-03 2.808 0.00541 **
Lu -3.418e-02 6.570e-03 -5.203 4.29e-07 ***
ETRLs -1.290e-03 6.411e-04 -2.012 0.04535 *
K -7.385e-05 3.045e-05 -2.425 0.01607 *
Al -7.018e-03 2.697e-03 -2.603 0.00984 **
pH 1.131e-02 2.617e-03 4.323 2.28e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2260052)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 71.037 on 233 degrees of freedom
AIC: 1713.7
Number of Fisher Scoring iterations: 5mlg6.11 = glm(formula = Argila ~ La + Ce + Tb + Lu +
ETRLs + K + Al + pH, family = Gamma(link = "inverse"),
data = dados[, c(-1, -3, -30)])
summary(mlg6.11)
Call:
glm(formula = Argila ~ La + Ce + Tb + Lu + ETRLs + K + Al + pH,
family = Gamma(link = "inverse"), data = dados[, c(-1, -3,
-30)])
Deviance Residuals:
Min 1Q Median 3Q Max
-2.65836 -0.33563 -0.05352 0.25311 1.51683
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 3.019e-02 1.147e-02 2.633 0.00903 **
La 1.309e-03 4.817e-04 2.717 0.00708 **
Ce 7.067e-04 2.662e-04 2.655 0.00847 **
Tb 2.041e-02 6.848e-03 2.981 0.00318 **
Lu -3.729e-02 5.749e-03 -6.486 5.19e-10 ***
ETRLs -7.380e-04 2.433e-04 -3.033 0.00270 **
K -7.757e-05 2.982e-05 -2.601 0.00989 **
Al -7.242e-03 2.658e-03 -2.725 0.00692 **
pH 1.102e-02 2.591e-03 4.252 3.07e-05 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Gamma family taken to be 0.2259654)
Null deviance: 94.641 on 242 degrees of freedom
Residual deviance: 71.227 on 234 degrees of freedom
AIC: 1712.3
Number of Fisher Scoring iterations: 5Verificando os resíduos do mlg6.11
par(mfrow = c(2,2))
plot(mlg6.11, pch = 20)
Envelope
set.seed(23445)
hnp(mlg6.11, print= T, main = "Gamma para var. Argila")Gamma model 