Duomenų gavyba Namų darbai
Antanas Kaminskas
library(stats)
library(ggplot2)
library(plot3D)
library(philentropy)## Warning: paketas 'philentropy' buvo sukurtas pagal R versiją 4.1.3library(factoextra)## Warning: paketas 'factoextra' buvo sukurtas pagal R versiją 4.1.3library(corrplot)
library(RColorBrewer)
library(PerformanceAnalytics)## Warning: paketas 'PerformanceAnalytics' buvo sukurtas pagal R versiją 4.1.3library(mlr)
library(tidyverse)
library(fastDummies)## Warning: paketas 'fastDummies' buvo sukurtas pagal R versiją 4.1.3library(e1071)
library(corrplot)
library(caTools)
library(class)
library(tree)## Warning: paketas 'tree' buvo sukurtas pagal R versiją 4.1.3library(readr)
library(tidyverse)
library(mlr)
library(scales)
library(HDclassif)
library(GGally)
library(purrr)
library(dplyr)Nuskaitomas pirmas duomenų rinkinys
Atribututų informacija
Attribute information:
For more information, read [Cortez and Morais, 2007].
- X - x- ašies specialiosios ilgio koordinatės Montesinho parke map: 1 to 9
- Y - y- ašies specialiosios pločio koordinatės Montesinho parke map: 2 to 9
- month - metų mėnesiai : “jan” to “dec”
- day - savaitės dienos: “mon” to “sun” perkoduota į darbo dienas ir savaitgalius
- FFMC - FFMC indeksas FWI sistemai: 18.7 to 96.20
- DMC - DMC indeksas FWI sistemai: 1.1 to 291.3
- DC - DC indeksas FWI sistemai: 7.9 to 860.6
- ISI - ISI indeksas FWI sistemai: 0.0 to 56.10
- temp - temperatūra Celsijais: 2.2 to 33.30
- RH - žmonių kaltės rodiklis %: 15.0 to 100
- wind - vėjo greitis km/h: 0.40 to 9.40
- rain - kritulių kiekis mm/m2 : 0.0 to 6.4
- area - išdegęs plotas hektarais: 0.00 to 1090.84
getwd()## [1] "C:/Users/antanas.kaminskas/Desktop"setwd("C:/Users/antanas.kaminskas/Desktop")data1 <- read.csv2("C:/Users/antanas.kaminskas/Desktop/DG_FF2.csv",
header = TRUE, sep = ";" ,dec = ".")
head(data1)## X Y day FFMC DMC DC ISI temp RH wind rain area
## 1 7 5 wd 86.2 26.2 94.3 5.1 8.2 51 6.7 0.0 0
## 2 7 4 wd 90.6 35.4 669.1 6.7 18.0 33 0.9 0.0 0
## 3 7 4 nwd 90.6 43.7 686.9 6.7 14.6 33 1.3 0.0 0
## 4 8 6 wd 91.7 33.3 77.5 9.0 8.3 97 4.0 0.2 0
## 5 8 6 nwd 89.3 51.3 102.2 9.6 11.4 99 1.8 0.0 0
## 6 8 6 nwd 92.3 85.3 488.0 14.7 22.2 29 5.4 0.0 0min.max.norm <- function(x, x.max, x.min)
{
return((x-x.min)/(x.max-x.min))
}
for(i in c(1,2,4,5,6))
{
max <- max(data1[,i])
min <- min(data1[,i])
for(ii in 1:nrow(data1))
{
data1[ii,i] <- min.max.norm(data1[ii,i], max, min)
}
}
ggplot(data1, aes(x=DC, fill=day)) +
geom_bar()
Aprašomoji statistika
summary(data1)## X Y day FFMC
## Min. :0.0000 Min. :0.0000 Length:517 Min. :0.0000
## 1st Qu.:0.2500 1st Qu.:0.2857 Class :character 1st Qu.:0.9226
## Median :0.3750 Median :0.2857 Mode :character Median :0.9406
## Mean :0.4587 Mean :0.3285 Mean :0.9283
## 3rd Qu.:0.7500 3rd Qu.:0.4286 3rd Qu.:0.9574
## Max. :1.0000 Max. :1.0000 Max. :1.0000
## DMC DC ISI temp
## Min. :0.0000 Min. :0.0000 Min. : 0.000 Min. : 2.20
## 1st Qu.:0.2326 1st Qu.:0.5040 1st Qu.: 6.500 1st Qu.:15.50
## Median :0.3694 Median :0.7697 Median : 8.400 Median :19.30
## Mean :0.3783 Mean :0.6333 Mean : 9.022 Mean :18.89
## 3rd Qu.:0.4869 3rd Qu.:0.8280 3rd Qu.:10.800 3rd Qu.:22.80
## Max. :1.0000 Max. :1.0000 Max. :56.100 Max. :33.30
## RH wind rain area
## Min. : 15.00 Min. :0.400 Min. :0.00000 Min. : 0.00
## 1st Qu.: 33.00 1st Qu.:2.700 1st Qu.:0.00000 1st Qu.: 0.00
## Median : 42.00 Median :4.000 Median :0.00000 Median : 0.52
## Mean : 44.29 Mean :4.018 Mean :0.02166 Mean : 12.85
## 3rd Qu.: 53.00 3rd Qu.:4.900 3rd Qu.:0.00000 3rd Qu.: 6.57
## Max. :100.00 Max. :9.400 Max. :6.40000 Max. :1090.84corrplot(cor(data1[,-3]), method = "number", type = "upper")
ggplot(data1, aes(x=as.factor(day) )) +
geom_bar(color="red", fill=rgb(0.1,0.4,0.5,0.7) )+ ggtitle("Day") +
xlab("Day") + ylab("Value")
data1_Task <- makeClassifTask(data = data1, target = "day")
lda <- makeLearner("classif.lda")
set.seed(50)
ldaModel <- train(lda, data1_Task)
ldaModelData <- getLearnerModel(ldaModel)
ldaPreds <- predict(ldaModelData)$x
head(ldaPreds)## LD1
## 1 0.4962508
## 2 1.1548667
## 3 1.8763079
## 4 -1.5180439
## 5 -2.5985542
## 6 0.6161613df <- cbind(data1, ldaPreds)
ggplot(df, aes(x=LD1, fill=DC, color=day)) +
geom_histogram()
qda <- makeLearner("classif.qda")
set.seed(50)
qdaModel <- train(qda, data1_Task)
kFold <- makeResampleDesc(method = "RepCV", folds = 3, reps = 10,
stratify = TRUE)
set.seed(50)
ldaCV <- resample(learner = lda, task = data1_Task, resampling = kFold,
measures = list(mmce, acc))
#qdaCV <- resample(learner = qda, task = data1_Task, resampling = kFold,
# measures = list(mmce, acc))
set.seed(50)
ldaCV$aggr## mmce.test.mean acc.test.mean
## 0.3446954 0.6553046#qdaCV$aggr
calculateConfusionMatrix(ldaCV$pred, relative = TRUE)## Relative confusion matrix (normalized by row/column):
## predicted
## true nwd wd -err.-
## nwd 0.10/0.51 0.90/0.33 0.90
## wd 0.05/0.49 0.95/0.67 0.05
## -err.- 0.49 0.33 0.34
##
##
## Absolute confusion matrix:
## predicted
## true nwd wd -err.-
## nwd 175 1615 1615
## wd 167 3213 167
## -err.- 167 1615 1782#calculateConfusionMatrix(qdaCV$pred, relative = TRUE)Modelio LDA tikslumas 65.53 %
QDA modelis neveikia, nes atsiranda klaida Error in checkClass(x, classes, ordered, null.ok) :
object ‘qdaCV’ not found
10-fold
kFold10 <- makeResampleDesc(method = "CV", iters = 10, stratify = TRUE)
set.seed(50)
ldaCVIA <- resample(learner = lda, task = data1_Task, resampling = kFold10, measures = list(mmce, acc))## Resampling: cross-validation## Measures: mmce acc## [Resample] iter 1: 0.3725490 0.6274510## [Resample] iter 2: 0.3461538 0.6538462## [Resample] iter 3: 0.3269231 0.6730769## [Resample] iter 4: 0.3461538 0.6538462## [Resample] iter 5: 0.3846154 0.6153846## [Resample] iter 6: 0.3461538 0.6538462## [Resample] iter 7: 0.3269231 0.6730769## [Resample] iter 8: 0.3653846 0.6346154## [Resample] iter 9: 0.3076923 0.6923077## [Resample] iter 10: 0.3000000 0.7000000## ## Aggregated Result: mmce.test.mean=0.3422549,acc.test.mean=0.6577451## ldaCVIA$aggr## mmce.test.mean acc.test.mean
## 0.3422549 0.6577451Modelio kFolds tikslumas 65,77 %
LOO
LOO <- makeResampleDesc(method = "LOO")
set.seed(50)
lda_LOO <- resample(learner = lda, task = data1_Task, resampling = LOO,
measures = list(mmce, acc))
lda_LOO$aggr## mmce.test.mean acc.test.mean
## 0.3462282 0.6537718LDA LOO tikslumas 65,54 %
KNN Hold out confusion matrix
set.seed(50)
dat.d <- sample(1:nrow(data1),size=nrow(data1)*0.7,replace = FALSE) #random selection of 70% data.
train.loan <- data1[dat.d,-3] # 70% training data
test.loan <- data1[-dat.d,-3] # remaining 30% test data
train.loan_labels <- data1[dat.d,3]
test.loan_labels <-data1[-dat.d,3]
i=1
k.optm=1
for (i in 1:12)
{
knn.mod <- knn(train=train.loan, test=test.loan, cl=train.loan_labels, k=i)
k.optm[i] <- 100 * sum(test.loan_labels == knn.mod)/NROW(test.loan_labels)
k=i
cat(k,'=',k.optm[i],''
)
}## 1 = 66.66667 2 = 57.69231 3 = 64.10256 4 = 56.41026 5 = 62.17949 6 = 64.74359 7 = 62.82051 8 = 62.17949 9 = 64.74359 10 = 64.10256 11 = 66.66667 12 = 67.30769plot(k.optm, type="b", xlab="K- Value",ylab="Accuracy level")
KNN atvaizduoja 12
knn <- makeLearner("classif.knn", par.vals = list("k" = 12))
holdoutNoStrat <- makeResampleDesc(method = "Holdout", split = 0.9, stratify = FALSE)
kFoldCV <- resample(learner = knn, task = data1_Task, resampling = holdoutNoStrat, measures = list(mmce, acc))## Resampling: holdout## Measures: mmce acc## [Resample] iter 1: 0.3653846 0.6346154## ## Aggregated Result: mmce.test.mean=0.3653846,acc.test.mean=0.6346154## KNN metodas atvaizduoja 63,46 %
Confusion Matrix:
calculateConfusionMatrix(kFoldCV$pred, relative = TRUE)## Relative confusion matrix (normalized by row/column):
## predicted
## true nwd wd -err.-
## nwd 0.12/0.29 0.88/0.31 0.88
## wd 0.14/0.71 0.86/0.69 0.14
## -err.- 0.71 0.31 0.37
##
##
## Absolute confusion matrix:
## predicted
## true nwd wd -err.-
## nwd 2 14 14
## wd 5 31 5
## -err.- 5 14 1910-fold crossvalidation
kFold10 <- makeResampleDesc(method = "CV", iters = 10, stratify = TRUE)
#IAModel <- train(IAda, IATask)
ldaCVIA <- resample(learner = knn, task = data1_Task, resampling = kFold10, measures = list(mmce, acc))## Resampling: cross-validation## Measures: mmce acc## [Resample] iter 1: 0.3269231 0.6730769## [Resample] iter 2: 0.4313725 0.5686275## [Resample] iter 3: 0.3269231 0.6730769## [Resample] iter 4: 0.3725490 0.6274510## [Resample] iter 5: 0.4230769 0.5769231## [Resample] iter 6: 0.4117647 0.5882353## [Resample] iter 7: 0.3846154 0.6153846## [Resample] iter 8: 0.3846154 0.6153846## [Resample] iter 9: 0.3461538 0.6538462## [Resample] iter 10: 0.4230769 0.5769231## ## Aggregated Result: mmce.test.mean=0.3831071,acc.test.mean=0.6168929## ldaCVIA$aggr## mmce.test.mean acc.test.mean
## 0.3831071 0.6168929KNN 10 -fold crossvalidation atvaizduoja 61,69 %
Logistic regression
Statistikoje logistinis modelis yra statistinis modelis, modeliuojantis vieno įvykio tikimybę, kai įvykio log-odds yra tiesinis vieno ar kelių nepriklausomų kintamųjų derinys. Regresinėje analizėje logistinė regresija yra logistinio modelio parametrų įvertinimas.
Hold out confusion matrix
data1_Task <- makeClassifTask(data = data1, target = "day")
logReg<-makeLearner("classif.logreg", predict.type="prob")
#logRegModel <- train(logReg, IA_Task)
logRegWrapper <- makeImputeWrapper("classif.logreg")
holdout <- makeResampleDesc(method = "Holdout", split = 0.9, stratify = TRUE)
irisLogReg <-resample(learner = logReg, task=data1_Task,
resampling = holdout,
measures = list(acc))## Resampling: holdout## Measures: acc## [Resample] iter 1: 0.6730769## ## Aggregated Result: acc.test.mean=0.6730769## calculateConfusionMatrix(irisLogReg$pred, relative = TRUE)## Relative confusion matrix (normalized by row/column):
## predicted
## true nwd wd -err.-
## nwd 0.06/1.00 0.94/0.33 0.94
## wd 0.00/0.00 1.00/0.67 0.00
## -err.- 0.00 0.33 0.33
##
##
## Absolute confusion matrix:
## predicted
## true nwd wd -err.-
## nwd 1 17 17
## wd 0 34 0
## -err.- 0 17 17Regrestion atvaizduoja 67,31 %
10-fold crossvalidation
kFold <- makeResampleDesc(method = "CV", iters = 10)
set.seed(50)
logRegwithImpute <- resample(logRegWrapper, data1_Task,
resampling = kFold,
measures = list(acc))## Resampling: cross-validation## Measures: acc## [Resample] iter 1: 0.6538462## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred## [Resample] iter 2: 0.7692308## [Resample] iter 3: 0.6730769## [Resample] iter 4: 0.5882353## [Resample] iter 5: 0.7115385## [Resample] iter 6: 0.6730769## [Resample] iter 7: 0.5490196## [Resample] iter 8: 0.6274510## [Resample] iter 9: 0.6538462## [Resample] iter 10: 0.6346154## ## Aggregated Result: acc.test.mean=0.6533937## calculateConfusionMatrix(logRegwithImpute$pred, relative = TRUE)## Relative confusion matrix (normalized by row/column):
## predicted
## true nwd wd -err.-
## nwd 0.09/0.50 0.91/0.34 0.91
## wd 0.05/0.50 0.95/0.66 0.05
## -err.- 0.50 0.34 0.35
##
##
## Absolute confusion matrix:
## predicted
## true nwd wd -err.-
## nwd 17 162 162
## wd 17 321 17
## -err.- 17 162 179Regression 10-fold 65,34 %
Principal components metodas
PC <- prcomp(data1[,-3], scale = TRUE)
PC## Standard deviations (1, .., p=11):
## [1] 1.6936493 1.2527005 1.1413274 1.1031844 0.9942302 0.9649257 0.8223573
## [8] 0.6893450 0.6804158 0.5420547 0.4634835
##
## Rotation (n x k) = (11 x 11):
## PC1 PC2 PC3 PC4 PC5 PC6
## X -0.07556259 0.678153236 -0.10331699 0.05522838 -0.07695738 0.06027244
## Y -0.06879035 0.669309745 -0.09826604 0.13918799 -0.11362354 0.08261238
## FFMC 0.42309818 0.053615017 -0.17567188 -0.25663655 -0.09165460 0.10704494
## DMC 0.42883564 0.110385775 0.43483082 0.03705563 0.09462031 0.15068344
## DC 0.43230750 0.001888655 0.38751278 0.18211511 0.05427566 0.07969746
## ISI 0.35802561 0.096922723 -0.12679514 -0.45672785 -0.09363709 0.20381273
## temp 0.48491200 0.018675558 -0.20732742 0.16337476 -0.04611169 -0.17995218
## RH -0.23301870 0.136933977 0.68739686 -0.14745909 0.06331626 0.19032738
## wind -0.12280037 0.023814794 -0.08727304 -0.69586380 0.25295051 0.21750241
## rain 0.04870842 0.175589517 0.22381117 -0.34454231 -0.02947805 -0.88589973
## area 0.06844108 0.131060942 -0.14241010 0.13116221 0.93867554 -0.08218313
## PC7 PC8 PC9 PC10 PC11
## X -0.01457914 0.39275003 0.58078880 -0.12904118 0.051829627
## Y -0.12517156 -0.34326853 -0.56717344 0.20545435 -0.033605505
## FFMC 0.34648591 0.59890885 -0.41707510 -0.01260941 -0.220627066
## DMC -0.20301350 -0.02228905 -0.15734444 -0.56407070 0.444028372
## DC -0.27361636 0.15313241 0.15261024 0.70608828 -0.017637296
## ISI 0.41893557 -0.50163283 0.28757769 0.16666976 0.223196190
## temp -0.23925105 -0.28574845 0.18166411 -0.28735575 -0.637069938
## RH 0.32136275 -0.09225706 0.02976710 -0.09814767 -0.519764225
## wind -0.60106243 0.04963205 -0.01587542 0.01499531 -0.130393154
## rain 0.01567295 0.01468460 -0.04354473 0.05242092 0.084866421
## area 0.22459490 -0.02824828 -0.01621226 0.03582682 -0.001268496PC$sdev## [1] 1.6936493 1.2527005 1.1413274 1.1031844 0.9942302 0.9649257 0.8223573
## [8] 0.6893450 0.6804158 0.5420547 0.4634835PC$rotation## PC1 PC2 PC3 PC4 PC5 PC6
## X -0.07556259 0.678153236 -0.10331699 0.05522838 -0.07695738 0.06027244
## Y -0.06879035 0.669309745 -0.09826604 0.13918799 -0.11362354 0.08261238
## FFMC 0.42309818 0.053615017 -0.17567188 -0.25663655 -0.09165460 0.10704494
## DMC 0.42883564 0.110385775 0.43483082 0.03705563 0.09462031 0.15068344
## DC 0.43230750 0.001888655 0.38751278 0.18211511 0.05427566 0.07969746
## ISI 0.35802561 0.096922723 -0.12679514 -0.45672785 -0.09363709 0.20381273
## temp 0.48491200 0.018675558 -0.20732742 0.16337476 -0.04611169 -0.17995218
## RH -0.23301870 0.136933977 0.68739686 -0.14745909 0.06331626 0.19032738
## wind -0.12280037 0.023814794 -0.08727304 -0.69586380 0.25295051 0.21750241
## rain 0.04870842 0.175589517 0.22381117 -0.34454231 -0.02947805 -0.88589973
## area 0.06844108 0.131060942 -0.14241010 0.13116221 0.93867554 -0.08218313
## PC7 PC8 PC9 PC10 PC11
## X -0.01457914 0.39275003 0.58078880 -0.12904118 0.051829627
## Y -0.12517156 -0.34326853 -0.56717344 0.20545435 -0.033605505
## FFMC 0.34648591 0.59890885 -0.41707510 -0.01260941 -0.220627066
## DMC -0.20301350 -0.02228905 -0.15734444 -0.56407070 0.444028372
## DC -0.27361636 0.15313241 0.15261024 0.70608828 -0.017637296
## ISI 0.41893557 -0.50163283 0.28757769 0.16666976 0.223196190
## temp -0.23925105 -0.28574845 0.18166411 -0.28735575 -0.637069938
## RH 0.32136275 -0.09225706 0.02976710 -0.09814767 -0.519764225
## wind -0.60106243 0.04963205 -0.01587542 0.01499531 -0.130393154
## rain 0.01567295 0.01468460 -0.04354473 0.05242092 0.084866421
## area 0.22459490 -0.02824828 -0.01621226 0.03582682 -0.001268496PC$center## X Y FFMC DMC DC ISI
## 0.45865571 0.32854380 0.92831846 0.37826444 0.63332947 9.02166344
## temp RH wind rain area
## 18.88916828 44.28820116 4.01760155 0.02166344 12.84729207PC$scale## X Y FFMC DMC DC ISI
## 0.28922223 0.17570006 0.07122724 0.22069773 0.29091848 4.55947718
## temp RH wind rain area
## 5.80662535 16.31746924 1.79165260 0.29595912 63.65581847head(PC$x)## PC1 PC2 PC3 PC4 PC5 PC6
## [1,] -3.3111267 0.8065462 -0.6469244 -1.05215166 0.09012872 0.3179227
## [2,] -0.2560234 0.1626998 -0.6168351 1.58821969 -0.75057626 -0.6038570
## [3,] -0.4807794 0.1715222 -0.4307642 1.35507053 -0.65094627 -0.4246828
## [4,] -3.0842255 2.2613875 1.1840076 -1.16676201 -0.42353215 0.3104463
## [5,] -2.7093024 2.1608966 1.3339775 0.06970309 -0.67154734 0.5998561
## [6,] 0.4758850 1.8551412 -1.5796407 -0.73950254 -0.54058097 0.3757952
## PC7 PC8 PC9 PC10 PC11
## [1,] -0.3302591 0.4658783 -0.05992081 -0.1917307 0.22176275
## [2,] 0.7190309 0.8567897 0.82587507 0.8194701 0.09473749
## [3,] 0.6789872 1.0432865 0.70651928 0.9686412 0.49493249
## [4,] 2.0794635 0.1814565 -0.38587063 -0.3296094 -0.98402165
## [5,] 2.5387721 -0.3706907 -0.04616543 -0.6096974 -1.03675031
## [6,] -0.3689603 -0.4159585 0.38094223 0.2852692 0.08410152eig.val <- get_eigenvalue(PC)
eig.val## eigenvalue variance.percent cumulative.variance.percent
## Dim.1 2.8684479 26.076799 26.07680
## Dim.2 1.5692585 14.265986 40.34279
## Dim.3 1.3026283 11.842075 52.18486
## Dim.4 1.2170159 11.063781 63.24864
## Dim.5 0.9884937 8.986306 72.23495
## Dim.6 0.9310817 8.464379 80.69933
## Dim.7 0.6762716 6.147924 86.84725
## Dim.8 0.4751966 4.319969 91.16722
## Dim.9 0.4629656 4.208779 95.37600
## Dim.10 0.2938233 2.671121 98.04712
## Dim.11 0.2148169 1.952881 100.00000res.var <- get_pca_var(PC)
res.var$coord ## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6
## X -0.12797653 0.849522889 -0.11791852 0.06092709 -0.07651335 0.05815842
## Y -0.11650672 0.838444643 -0.11215372 0.15355003 -0.11296795 0.07971481
## FFMC 0.71657993 0.067163557 -0.20049913 -0.28311745 -0.09112578 0.10329042
## DMC 0.72629718 0.138280314 0.49628434 0.04087919 0.09407437 0.14539833
## DC 0.73217729 0.002365919 0.44227896 0.20090655 0.05396250 0.07690213
## ISI 0.60636982 0.121415142 -0.14471477 -0.50385506 -0.09309682 0.19666415
## temp 0.82127086 0.023394880 -0.23662847 0.18023250 -0.04584563 -0.17364049
## RH -0.39465196 0.171537259 0.78454489 -0.16267457 0.06295094 0.18365179
## wind -0.20798076 0.029832804 -0.09960712 -0.76766612 0.25149104 0.20987368
## rain 0.08249498 0.219961074 0.25544182 -0.38009371 -0.02930796 -0.85482745
## area 0.11591518 0.164180106 -0.16253656 0.14469611 0.93325958 -0.07930061
## Dim.7 Dim.8 Dim.9 Dim.10 Dim.11
## X -0.01198926 0.27074028 0.39517786 -0.069947383 0.0240221750
## Y -0.10293575 -0.23663045 -0.38591376 0.111367498 -0.0155755959
## FFMC 0.28493524 0.41285484 -0.28378448 -0.006834990 -0.1022569972
## DMC -0.16694964 -0.01536484 -0.10705964 -0.305757184 0.2057998088
## DC -0.22501042 0.10556106 0.10383841 0.382738483 -0.0081745950
## ISI 0.34451474 -0.34579809 0.19567240 0.090344132 0.1034477434
## temp -0.19674986 -0.19697927 0.12360713 -0.155762542 -0.2952713828
## RH 0.26427502 -0.06359695 0.02025400 -0.053201406 -0.2409021240
## wind -0.49428810 0.03421360 -0.01080188 0.008128276 -0.0604350710
## rain 0.01288877 0.01012276 -0.02962852 0.028415009 0.0393341827
## area 0.18469727 -0.01947281 -0.01103107 0.019420096 -0.0005879271res.var$contrib ## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6
## X 0.5709706 4.598918e+01 1.0674401 0.3050174 0.59224380 0.3632766
## Y 0.4732112 4.479755e+01 0.9656214 1.9373297 1.29103080 0.6824806
## FFMC 17.9012071 2.874570e-01 3.0860609 6.5862319 0.84005666 1.1458620
## DMC 18.3900010 1.218502e+00 18.9077845 0.1373119 0.89530033 2.2705499
## DC 18.6889775 3.567017e-04 15.0166153 3.3165912 0.29458473 0.6351685
## ISI 12.8182340 9.394014e-01 1.6077007 20.8600325 0.87679048 4.1539630
## temp 23.5139649 3.487765e-02 4.2984660 2.6691314 0.21262877 3.2382786
## RH 5.4297716 1.875091e+00 47.2514450 2.1744182 0.40089488 3.6224511
## wind 1.5079931 5.671444e-02 0.7616584 48.4226431 6.39839626 4.7307300
## rain 0.2372510 3.083168e+00 5.0091439 11.8709401 0.08689552 78.4818331
## area 0.4684181 1.717697e+00 2.0280638 1.7203525 88.11117777 0.6754066
## Dim.7 Dim.8 Dim.9 Dim.10 Dim.11
## X 0.02125514 15.42525895 33.73156252 1.66516274 2.686310e-01
## Y 1.56679190 11.78332820 32.16857099 4.22114883 1.129330e-01
## FFMC 12.00524890 35.86918138 17.39516426 0.01589972 4.867630e+00
## DMC 4.12144801 0.04968015 2.47572732 31.81757538 1.971612e+01
## DC 7.48659109 2.34495341 2.32898848 49.85606565 3.110742e-02
## ISI 17.55070088 25.16354950 8.27009262 2.77788102 4.981654e+00
## temp 5.72410665 8.16521762 3.30018489 8.25733288 4.058581e+01
## RH 10.32740197 0.85113651 0.08860800 0.96329644 2.701548e+01
## wind 36.12760443 0.24633400 0.02520289 0.02248592 1.700237e+00
## rain 0.02456415 0.02156375 0.18961431 0.27479533 7.202309e-01
## area 5.04428690 0.07979654 0.02628372 0.12835609 1.609083e-04res.var$cos2 ## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6
## X 0.016377993 7.216891e-01 0.013904776 0.003712110 0.0058542927 0.003382402
## Y 0.013573816 7.029894e-01 0.012578457 0.023577611 0.0127617581 0.006354452
## FFMC 0.513486795 4.510943e-03 0.040199902 0.080155491 0.0083039071 0.010668911
## DMC 0.527507595 1.912145e-02 0.246298146 0.001671108 0.0088499873 0.021140674
## DC 0.536083580 5.597571e-06 0.195610676 0.040363443 0.0029119515 0.005913937
## ISI 0.367684362 1.474164e-02 0.020942363 0.253869917 0.0086670186 0.038676788
## temp 0.674485830 5.473204e-04 0.055993033 0.032483754 0.0021018219 0.030151019
## RH 0.155750169 2.942503e-02 0.615510679 0.026463016 0.0039628206 0.033727978
## wind 0.043255997 8.899962e-04 0.009921577 0.589311277 0.0632477438 0.044046959
## rain 0.006805421 4.838287e-02 0.065250525 0.144471231 0.0008589568 0.730729961
## area 0.013436329 2.695511e-02 0.026418132 0.020936964 0.8709734397 0.006288587
## Dim.7 Dim.8 Dim.9 Dim.10 Dim.11
## X 0.0001437424 0.0733003004 0.1561655432 4.892636e-03 5.770649e-04
## Y 0.0105957688 0.0559939706 0.1489294295 1.240272e-02 2.425992e-04
## FFMC 0.0811880895 0.1704491171 0.0805336330 4.671709e-05 1.045649e-02
## DMC 0.0278721826 0.0002360784 0.0114617668 9.348746e-02 4.235356e-02
## DC 0.0506296898 0.0111431380 0.0107824164 1.464887e-01 6.682400e-05
## ISI 0.1186904066 0.1195763224 0.0382876870 8.162062e-03 1.070144e-02
## temp 0.0387105080 0.0388008335 0.0152787220 2.426197e-02 8.718519e-02
## RH 0.0698412871 0.0040445714 0.0004102246 2.830390e-03 5.803383e-02
## wind 0.2443207305 0.0011705707 0.0001166807 6.606887e-05 3.652398e-03
## rain 0.0001661203 0.0001024702 0.0008778491 8.074127e-04 1.547178e-03
## area 0.0341130800 0.0003791904 0.0001216846 3.771401e-04 3.456583e-07res.ind <- get_pca_ind(PC)
head(res.ind$coord) ## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6 Dim.7
## 1 -3.3111267 0.8065462 -0.6469244 -1.05215166 0.09012872 0.3179227 -0.3302591
## 2 -0.2560234 0.1626998 -0.6168351 1.58821969 -0.75057626 -0.6038570 0.7190309
## 3 -0.4807794 0.1715222 -0.4307642 1.35507053 -0.65094627 -0.4246828 0.6789872
## 4 -3.0842255 2.2613875 1.1840076 -1.16676201 -0.42353215 0.3104463 2.0794635
## 5 -2.7093024 2.1608966 1.3339775 0.06970309 -0.67154734 0.5998561 2.5387721
## 6 0.4758850 1.8551412 -1.5796407 -0.73950254 -0.54058097 0.3757952 -0.3689603
## Dim.8 Dim.9 Dim.10 Dim.11
## 1 0.4658783 -0.05992081 -0.1917307 0.22176275
## 2 0.8567897 0.82587507 0.8194701 0.09473749
## 3 1.0432865 0.70651928 0.9686412 0.49493249
## 4 0.1814565 -0.38587063 -0.3296094 -0.98402165
## 5 -0.3706907 -0.04616543 -0.6096974 -1.03675031
## 6 -0.4159585 0.38094223 0.2852692 0.08410152head(res.ind$contrib) ## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6
## 1 0.739288748 0.080181376 0.06214354 0.1759421474 0.001589505 0.02099736
## 2 0.004419995 0.003262791 0.05649723 0.4008984211 0.110236444 0.07575127
## 3 0.015586701 0.003626233 0.02755295 0.2918347448 0.082913614 0.03746719
## 4 0.641438091 0.630325592 0.20816031 0.2163603915 0.035100097 0.02002140
## 5 0.494968370 0.575549858 0.26423235 0.0007721778 0.088244738 0.07475080
## 6 0.015270967 0.424198284 0.37051485 0.0869145609 0.057181699 0.02933757
## Dim.7 Dim.8 Dim.9 Dim.10 Dim.11
## 1 0.03119592 0.08834484 0.0015000857 0.02419950 0.044281076
## 2 0.14787113 0.29880279 0.2849636180 0.44206839 0.008081365
## 3 0.13185951 0.44304043 0.2085492971 0.61765948 0.220563115
## 4 1.23677556 0.01340236 0.0622077014 0.07151915 0.871867061
## 5 1.84346859 0.05593189 0.0008904192 0.24471012 0.967808244
## 6 0.03893566 0.07042651 0.0606287946 0.05357139 0.006368668head(res.ind$cos2)## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6
## 1 0.80234169 0.047606505 0.03062773 0.0810148154 0.0005944757 0.007396918
## 2 0.01002780 0.004049680 0.05820826 0.3858936223 0.0861858748 0.055784647
## 3 0.03777996 0.004808509 0.03032836 0.3001190227 0.0692564164 0.029478078
## 4 0.40916890 0.219968421 0.06030028 0.0585564750 0.0077158385 0.004145560
## 5 0.32427004 0.206281298 0.07861201 0.0002146327 0.0199225596 0.015895934
## 6 0.02946360 0.447750125 0.32463731 0.0711477413 0.0380192267 0.018373177
## Dim.7 Dim.8 Dim.9 Dim.10 Dim.11
## 1 0.007982102 0.015883741 2.627623e-04 0.002690240 0.0035990257
## 2 0.079093634 0.112303923 1.043458e-01 0.102733649 0.0013730622
## 3 0.075351667 0.177900399 8.158640e-02 0.153354186 0.0400370107
## 4 0.186000004 0.001416301 6.404627e-03 0.004673148 0.0416504470
## 5 0.284733987 0.006070370 9.415117e-05 0.016421794 0.0474832264
## 6 0.017710914 0.022510325 1.887991e-02 0.010587454 0.0009202153fviz_eig(PC)
fviz_pca_ind(PC,
col.ind = "cos2", # Color by the quality of representation
gradient.cols = c("#00AFBB", "#E7B800", "#FC4E07"),
repel = TRUE # Avoid text overlapping
)## Warning: ggrepel: 474 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps
fviz_pca_var(PC,
col.var = "contrib", # Color by contributions to the PC
gradient.cols = c("#00AFBB", "#E7B800", "#FC4E07"),
repel = TRUE # Avoid text overlapping
)
PC atvaizduoja 44,4 %
Išvados
- LDA modelis atvaizduoja 65,53 % duomenų
- LDA 10-fold modelis atvaizduoja 65,77 % duomenų
- LDA LOO modelis atvaizduoja 65,54 % duomenų
- KNN modelis atvaizduoja 63,46 % duomenų
- KNN 10-fold modelis atvaizduoja 61,69 % duomenų
- Regression modelis atvaizduoja 67,31 % duomenų
- Regression modelis atvaizduoja 63,46 % duomenų
- PC pirmosios dvi komponentės modelis atvaizduoja 44,4 % duomenų