#### TIỀN XỬ LÝ DỮ LIỆU:
# 1. ĐỌC DỮ LIỆU & XEM TRƯỚC
# Đọc dữ liệu từ file CSV
printer_data <- read.csv("D:/HCMUT/HK252/BTL XSTK/data.csv")
# Xem 10 dòng đầu tiên
head(printer_data,10)
## layer_height wall_thickness infill_density infill_pattern nozzle_temperature
## 1 0.02 8 90 grid 220
## 2 0.02 7 90 honeycomb 225
## 3 0.02 1 80 grid 230
## 4 0.02 4 70 honeycomb 240
## 5 0.02 6 90 grid 250
## 6 0.02 10 40 honeycomb 200
## 7 0.02 5 10 grid 205
## 8 0.02 10 10 honeycomb 210
## 9 0.02 9 70 grid 215
## 10 0.02 8 40 honeycomb 220
## bed_temperature print_speed material fan_speed roughness tension_strenght
## 1 60 40 abs 0 25 18
## 2 65 40 abs 25 32 16
## 3 70 40 abs 50 40 8
## 4 75 40 abs 75 68 10
## 5 80 40 abs 100 92 5
## 6 60 40 pla 0 60 24
## 7 65 40 pla 25 55 12
## 8 70 40 pla 50 21 14
## 9 75 40 pla 75 24 27
## 10 80 40 pla 100 30 25
## elongation
## 1 1.2
## 2 1.4
## 3 0.8
## 4 0.5
## 5 0.7
## 6 1.1
## 7 1.3
## 8 1.5
## 9 1.4
## 10 1.7
# 2. CHỌN BIẾN CẦN PHÂN TÍCH
# Lấy các biến liên quan đến độ nhám (roughness)
new_DF <- printer_data[,c("layer_height","wall_thickness","infill_density",
"infill_pattern","nozzle_temperature","bed_temperature","print_speed",
"material","fan_speed","roughness")]
head(new_DF,10)
## layer_height wall_thickness infill_density infill_pattern nozzle_temperature
## 1 0.02 8 90 grid 220
## 2 0.02 7 90 honeycomb 225
## 3 0.02 1 80 grid 230
## 4 0.02 4 70 honeycomb 240
## 5 0.02 6 90 grid 250
## 6 0.02 10 40 honeycomb 200
## 7 0.02 5 10 grid 205
## 8 0.02 10 10 honeycomb 210
## 9 0.02 9 70 grid 215
## 10 0.02 8 40 honeycomb 220
## bed_temperature print_speed material fan_speed roughness
## 1 60 40 abs 0 25
## 2 65 40 abs 25 32
## 3 70 40 abs 50 40
## 4 75 40 abs 75 68
## 5 80 40 abs 100 92
## 6 60 40 pla 0 60
## 7 65 40 pla 25 55
## 8 70 40 pla 50 21
## 9 75 40 pla 75 24
## 10 80 40 pla 100 30
# 3. KIỂM TRA DỮ LIỆU KHUYẾT
# Kiểm tra giá trị thiếu (NA)
colSums(is.na(new_DF))
## layer_height wall_thickness infill_density infill_pattern
## 0 0 0 0
## nozzle_temperature bed_temperature print_speed material
## 0 0 0 0
## fan_speed roughness
## 0 0
# 4. KIỂM TRA NGOẠI LAI (OUTLIERS)
# Chọn các biến liên tục
continous_data <- new_DF[,c("layer_height","wall_thickness","infill_density",
"nozzle_temperature","bed_temperature","print_speed",
"fan_speed","roughness")]
# Dùng IQR để phát hiện ngoại lai
sapply(continous_data, function(x) {
out <- x < quantile(x,0.25) - 1.5*IQR(x) |
x > quantile(x,0.75) + 1.5*IQR(x)
c(count = sum(out), # số lượng outlier
rate = round(mean(out)*100,2)) # tỷ lệ %
})
## layer_height wall_thickness infill_density nozzle_temperature
## count 0 0 0 0
## rate 0 0 0 0
## bed_temperature print_speed fan_speed roughness
## count 0 10 0 0
## rate 0 20 0 0
# Kiểm tra dữ liệu print_speed
table(new_DF$print_speed)
##
## 40 60 120
## 20 20 10
# 5. KIỂM TRA DỮ LIỆU TRÙNG LẶP
# Kiểm tra bản ghi trùng lặp
sum(duplicated(new_DF))
## [1] 0
# 6. KHAI BÁO BIẾN PHÂN LOẠI
# Chuyển biến phân loại sang factor
new_DF$infill_pattern <- as.factor(new_DF$infill_pattern)
new_DF$material <- as.factor(new_DF$material)
#### THỐNG KÊ MÔ TẢ
# 1. TÍNH TOÁN THỐNG KÊ MÔ TẢ
# Tính toán hống kê mô tả cho dữ liệu
summary(new_DF)
## layer_height wall_thickness infill_density infill_pattern
## Min. :0.020 Min. : 1.00 Min. :10.0 grid :25
## 1st Qu.:0.060 1st Qu.: 3.00 1st Qu.:40.0 honeycomb:25
## Median :0.100 Median : 5.00 Median :50.0
## Mean :0.106 Mean : 5.22 Mean :53.4
## 3rd Qu.:0.150 3rd Qu.: 7.00 3rd Qu.:80.0
## Max. :0.200 Max. :10.00 Max. :90.0
## nozzle_temperature bed_temperature print_speed material fan_speed
## Min. :200.0 Min. :60 Min. : 40 abs:25 Min. : 0
## 1st Qu.:210.0 1st Qu.:65 1st Qu.: 40 pla:25 1st Qu.: 25
## Median :220.0 Median :70 Median : 60 Median : 50
## Mean :221.5 Mean :70 Mean : 64 Mean : 50
## 3rd Qu.:230.0 3rd Qu.:75 3rd Qu.: 60 3rd Qu.: 75
## Max. :250.0 Max. :80 Max. :120 Max. :100
## roughness
## Min. : 21.0
## 1st Qu.: 92.0
## Median :165.5
## Mean :170.6
## 3rd Qu.:239.2
## Max. :368.0
# 2. TRỰC QUAN HÓA BẰNG CÁC ĐỒ THỊ
library(ggplot2)
library(reshape2)
# Histogram roughness
ggplot(new_DF, aes(x = roughness)) +
geom_histogram(fill = "pink", bins = 8, color = "black") +
labs(title = "Histogram of Roughness",
x = "roughness", y = "Frequency") +
theme_minimal()
# Boxplot: roughness theo infill_pattern
ggplot(new_DF, aes(x = infill_pattern, y = roughness, fill = infill_pattern)) +
geom_boxplot() +
labs(title = "Roughness vs Infill Pattern",
x = "Infill Pattern", y = "Roughness") +
theme_minimal()
# Boxplot: roughness theo material
ggplot(new_DF, aes(x = material, y = roughness, fill = material)) +
geom_boxplot() +
labs(title = "Roughness vs Material",
x = "Material", y = "Roughness") +
theme_minimal()
# Hàm vẽ scatter + regression line
plot_scatter <- function(x_var){
ggplot(new_DF, aes_string(x = x_var, y = "roughness")) +
geom_point(color = "blue", alpha = 0.6) +
geom_smooth(method = "lm", se = FALSE, color = "red") +
labs(title = paste(x_var, "& Roughness"),
x = x_var, y = "Roughness") +
theme_minimal()
}
# Vẽ scatter cho từng biến
plot_scatter("layer_height")
## Warning: `aes_string()` was deprecated in ggplot2 3.0.0.
## ℹ Please use tidy evaluation idioms with `aes()`.
## ℹ See also `vignette("ggplot2-in-packages")` for more information.
## This warning is displayed once per session.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.
## `geom_smooth()` using formula = 'y ~ x'
plot_scatter("wall_thickness")
## `geom_smooth()` using formula = 'y ~ x'
plot_scatter("infill_density")
## `geom_smooth()` using formula = 'y ~ x'
plot_scatter("nozzle_temperature")
## `geom_smooth()` using formula = 'y ~ x'
plot_scatter("bed_temperature")
## `geom_smooth()` using formula = 'y ~ x'
plot_scatter("print_speed")
## `geom_smooth()` using formula = 'y ~ x'
plot_scatter("fan_speed")
## `geom_smooth()` using formula = 'y ~ x'
# Vẽ ma trận tương quan
cor_matrix <- cor(continous_data)
cor_melt <- melt(cor_matrix)
ggplot(cor_melt, aes(x = Var1, y = Var2, fill = value)) +
geom_tile() +
geom_text(aes(label = round(value, 2)), size = 3) +
scale_fill_gradient2(low = "red", high = "blue", mid = "white", midpoint = 0) +
labs(title = "Correlation Matrix") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
#### THỐNG KÊ SUY DIỄN:
# 1. KIỂM ĐỊNH TRUNG BÌNH 2 MẪU: (ABS vs PLA)
# Tách dữ liệu theo vật liệu
abs_data <- subset(new_DF, material=="abs")
pla_data <- subset(new_DF, material=="pla")
# QQ-plot kiểm tra phân phối chuẩn
par(mfrow=c(1,2))
qqnorm(abs_data$roughness,main="QQ-Plot - ABS")
qqline(abs_data$roughness,col="red")
qqnorm(pla_data$roughness,main="QQ-Plot - PLA")
qqline(pla_data$roughness,col="red")
# Shapiro test
by(new_DF$roughness,new_DF$material,shapiro.test)
## new_DF$material: abs
##
## Shapiro-Wilk normality test
##
## data: dd[x, ]
## W = 0.94235, p-value = 0.1677
##
## ------------------------------------------------------------
## new_DF$material: pla
##
## Shapiro-Wilk normality test
##
## data: dd[x, ]
## W = 0.97437, p-value = 0.7561
# Kiểm định phương sai
var.test(roughness~material,data=new_DF)
##
## F test to compare two variances
##
## data: roughness by material
## F = 1.8454, num df = 24, denom df = 24, p-value = 0.1405
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
## 0.8132207 4.1877771
## sample estimates:
## ratio of variances
## 1.845423
# T-test
t.test(roughness~material,var.equal=TRUE,data=new_DF)
##
## Two Sample t-test
##
## data: roughness by material
## t = 1.6613, df = 48, p-value = 0.1032
## alternative hypothesis: true difference in means between group abs and group pla is not equal to 0
## 95 percent confidence interval:
## -9.615162 101.055162
## sample estimates:
## mean in group abs mean in group pla
## 193.44 147.72
# 2. ANOVA 2 YẾU TỐ
# Tạo nhóm nhiệt độ nozzle
new_DF$nozzle_temperature_group <- cut(new_DF$nozzle_temperature,
breaks = c(-Inf,210,225, Inf),
labels = c("Low", "Medium", "High"))
# Chuyển sang factor
new_DF$layer_height <- as.factor(new_DF$layer_height)
new_DF$nozzle_temperature_group <- as.factor(new_DF$nozzle_temperature_group)
# Bảng chéo
table(new_DF$layer_height, new_DF$nozzle_temperature_group)
##
## Low Medium High
## 0.02 3 4 3
## 0.06 3 4 3
## 0.1 3 4 3
## 0.15 3 4 3
## 0.2 3 4 3
# Mô hình ANOVA 2 yếu tố
model_anova <- aov(roughness ~ layer_height*nozzle_temperature_group, data = new_DF)
summary(model_anova)
## Df Sum Sq Mean Sq F value Pr(>F)
## layer_height 4 326957 81739 93.549 < 2e-16 ***
## nozzle_temperature_group 2 100223 50112 57.352 9.01e-12 ***
## layer_height:nozzle_temperature_group 8 22819 2852 3.264 0.00697 **
## Residuals 35 30582 874
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# Tukey test
TukeyHSD(model_anova)$layer_height
## diff lwr upr p adj
## 0.06-0.02 78.3 40.29344 116.30656 9.295319e-06
## 0.1-0.02 150.5 112.49344 188.50656 2.413514e-12
## 0.15-0.02 162.2 124.19344 200.20656 1.739719e-13
## 0.2-0.02 238.4 200.39344 276.40656 0.000000e+00
## 0.1-0.06 72.2 34.19344 110.20656 3.730628e-05
## 0.15-0.06 83.9 45.89344 121.90656 2.599693e-06
## 0.2-0.06 160.1 122.09344 198.10656 3.145262e-13
## 0.15-0.1 11.7 -26.30656 49.70656 9.004931e-01
## 0.2-0.1 87.9 49.89344 125.90656 1.050935e-06
## 0.2-0.15 76.2 38.19344 114.20656 1.500107e-05
TukeyHSD(model_anova)$nozzle_temperature_group
## diff lwr upr p adj
## Medium-Low -32.20000 -56.90887 -7.491134 8.238541e-03
## High-Low 74.86667 48.45178 101.281555 1.370368e-07
## High-Medium 107.06667 82.35780 131.775533 5.212164e-12
# Interaction plot
par(mfrow=c(1,1))
interaction.plot(x.factor = new_DF$layer_height,
trace.factor = new_DF$nozzle_temperature_group,
response = new_DF$roughness,
col = c("red", "blue"),
lwd = 2)
# Kiểm tra các giả định của ANOVA
par(mfrow=c(1,2))
res <- residuals(model_anova)
hist(res)
qqnorm(res)
qqline(res, col="red")
# Normality
shapiro.test(res)
##
## Shapiro-Wilk normality test
##
## data: res
## W = 0.9824, p-value = 0.6566
# Homogeneity
library(car)
## Loading required package: carData
leveneTest(roughness ~ layer_height*nozzle_temperature_group, data = new_DF)
## Levene's Test for Homogeneity of Variance (center = median)
## Df F value Pr(>F)
## group 14 0.752 0.7097
## 35
# 3. HỒI QUY TUYẾN TÍNH
# Chuyển lại numeric
new_DF$layer_height <- as.numeric(as.character(new_DF$layer_height))
new_DF$nozzle_temperature_group <- NULL
# Model đầy đủ
model_1 <- lm(roughness~layer_height+nozzle_temperature+wall_thickness+infill_density+
infill_pattern+bed_temperature+print_speed+material,data=new_DF)
summary(model_1)
##
## Call:
## lm(formula = roughness ~ layer_height + nozzle_temperature +
## wall_thickness + infill_density + infill_pattern + bed_temperature +
## print_speed + material, data = new_DF)
##
## Residuals:
## Min 1Q Median 3Q Max
## -72.746 -24.332 -1.641 20.304 96.552
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2.371e+03 3.716e+02 -6.379 1.25e-07 ***
## layer_height 1.269e+03 8.765e+01 14.483 < 2e-16 ***
## nozzle_temperature 1.506e+01 2.529e+00 5.953 5.05e-07 ***
## wall_thickness 2.334e+00 2.189e+00 1.066 0.29259
## infill_density -4.231e-02 2.341e-01 -0.181 0.85742
## infill_patternhoneycomb -1.255e-01 1.128e+01 -0.011 0.99117
## bed_temperature -1.613e+01 3.251e+00 -4.962 1.27e-05 ***
## print_speed 6.496e-01 2.060e-01 3.153 0.00302 **
## materialpla 2.985e+02 5.836e+01 5.114 7.78e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 38.24 on 41 degrees of freedom
## Multiple R-squared: 0.8752, Adjusted R-squared: 0.8509
## F-statistic: 35.95 on 8 and 41 DF, p-value: 3.834e-16
# Model rút gọn
model_2 <- lm(roughness~layer_height+nozzle_temperature+
bed_temperature+print_speed+material,data=new_DF)
summary(model_2)
##
## Call:
## lm(formula = roughness ~ layer_height + nozzle_temperature +
## bed_temperature + print_speed + material, data = new_DF)
##
## Residuals:
## Min 1Q Median 3Q Max
## -74.084 -26.500 -1.662 22.585 92.356
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2310.7356 353.2009 -6.542 5.38e-08 ***
## layer_height 1246.5353 83.1780 14.986 < 2e-16 ***
## nozzle_temperature 14.7774 2.3979 6.163 1.95e-07 ***
## bed_temperature -15.8078 3.0895 -5.117 6.55e-06 ***
## print_speed 0.5538 0.1804 3.070 0.00366 **
## materialpla 294.1610 56.1586 5.238 4.38e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 37.44 on 44 degrees of freedom
## Multiple R-squared: 0.8717, Adjusted R-squared: 0.8571
## F-statistic: 59.78 on 5 and 44 DF, p-value: < 2.2e-16
# Kiểm tra đa cộng tuyến
library(car)
vif(model_2)
## layer_height nozzle_temperature bed_temperature print_speed
## 1.00309 44.15385 17.02564 1.00309
## material
## 28.12821
# Model sau khi xử lý đa cộng tuyến
model_3 <- lm(roughness~layer_height+nozzle_temperature+print_speed,data=new_DF)
summary(model_3)
##
## Call:
## lm(formula = roughness ~ layer_height + nozzle_temperature +
## print_speed, data = new_DF)
##
## Residuals:
## Min 1Q Median 3Q Max
## -77.37 -31.87 -8.78 34.56 107.49
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -512.9882 101.6645 -5.046 7.53e-06 ***
## layer_height 1246.5353 103.7800 12.011 8.78e-16 ***
## nozzle_temperature 2.3295 0.4502 5.174 4.89e-06 ***
## print_speed 0.5538 0.2251 2.461 0.0177 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 46.71 on 46 degrees of freedom
## Multiple R-squared: 0.7912, Adjusted R-squared: 0.7775
## F-statistic: 58.09 on 3 and 46 DF, p-value: 1.115e-15
vif(model_3)
## layer_height nozzle_temperature print_speed
## 1.00309 1.00000 1.00309
# Kiểm tra giả định
par(mfrow=c(2,2))
plot(model_3)
# Durbin-Watson test
library(lmtest)
## Loading required package: zoo
##
## Attaching package: 'zoo'
##
## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numeric
dwtest(model_3)
##
## Durbin-Watson test
##
## data: model_3
## DW = 0.74163, p-value = 3.386e-08
## alternative hypothesis: true autocorrelation is greater than 0
#### ĐỀ XUẤT MỞ RỘNG:
# 1 RESPONSE SURFACE METHOD (RSM)
library(rsm)
# Chuẩn hóa biến
m1 <- mean(new_DF$layer_height); s1 <- sd(new_DF$layer_height)
m2 <- mean(new_DF$nozzle_temperature); s2 <- sd(new_DF$nozzle_temperature)
m3 <- mean(new_DF$print_speed); s3 <- sd(new_DF$print_speed)
# Code dữ liệu
coded_dat <- coded.data(new_DF,
x1 ~ (layer_height - m1)/s1,
x2 ~ (nozzle_temperature - m2)/s2,
x3 ~ (print_speed - m3)/s3)
# Mô hình RSM
model_rsm <- rsm(roughness ~ FO(x1, x2, x3) +
TWI(x1, x2, x3) +
PQ(x1, x2),
data = coded_dat)
summary(model_rsm)
##
## Call:
## rsm(formula = roughness ~ FO(x1, x2, x3) + TWI(x1, x2, x3) +
## PQ(x1, x2), data = coded_dat)
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 137.5916 9.3565 14.7055 < 2.2e-16 ***
## x1 55.0801 10.3714 5.3108 4.107e-06 ***
## x2 21.4289 4.1315 5.1867 6.139e-06 ***
## x3 14.1869 6.9602 2.0383 0.04801 *
## x1:x2 5.3712 3.8686 1.3884 0.17251
## x1:x3 -37.1380 14.5307 -2.5558 0.01440 *
## x2:x3 17.5586 3.8681 4.5394 4.867e-05 ***
## x1^2 1.5581 8.1071 0.1922 0.84854
## x2^2 30.0410 3.5903 8.3672 2.086e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Multiple R-squared: 0.9389, Adjusted R-squared: 0.927
## F-statistic: 78.73 on 8 and 41 DF, p-value: < 2.2e-16
##
## Analysis of Variance Table
##
## Response: roughness
## Df Sum Sq Mean Sq F value Pr(>F)
## FO(x1, x2, x3) 3 380218 126739 176.9096 < 2.2e-16
## TWI(x1, x2, x3) 3 20807 6936 9.6812 5.897e-05
## PQ(x1, x2) 2 50182 25091 35.0235 1.346e-09
## Residuals 41 29373 716
## Lack of fit 36 29315 814 70.1981 7.737e-05
## Pure error 5 58 12
##
## Stationary point of response surface:
## x1 x2 x3
## 0.0224423 -0.7605069 1.3750121
##
## Stationary point in original units:
## layer_height nozzle_temperature print_speed
## 0.1074453 210.2292881 104.8241106
##
## Eigenanalysis:
## eigen() decomposition
## $values
## [1] 32.65982 18.12250 -19.18317
##
## $vectors
## [,1] [,2] [,3]
## x1 -0.1059075 -0.7351160 0.6696178
## x2 0.9435473 -0.2868364 -0.1656607
## x3 0.3138506 0.6142714 0.7239948
# Hướng tối ưu
st <- steepest(model_rsm)
## Path of steepest ascent from ridge analysis:
# Tìm giá trị độ nhám tối ưu
results <- data.frame()
for(i in 1:nrow(st)){
# chuyển về giá trị thực
layer_real <- st$x1[i]*s1 + m1
temp_real <- st$x2[i]*s2 + m2
speed_real <- st$x3[i]*s3 + m3
# dự đoán
pred <- predict(model_3,
newdata = data.frame(
layer_height = layer_real,
nozzle_temperature = temp_real,
print_speed = speed_real
))
results <- rbind(results,
data.frame(dist = st$dist[i],
layer = layer_real,
temp = temp_real,
speed = speed_real,
roughness = pred))
}
# Kết quả tối ưu
results
## dist layer temp speed roughness
## 1 0.0 0.1060000 221.5000 64.00000 170.5800
## 11 0.5 0.1318231 225.9016 64.95015 213.5494
## 12 1.0 0.1470851 232.8969 65.27677 249.0504
## 13 1.5 0.1535892 240.7665 67.50369 276.7238
## 14 2.0 0.1555211 248.6361 70.91831 299.3555
## 15 2.5 0.1553923 256.3426 74.86738 319.3348
## 16 3.0 0.1541688 263.9010 79.08369 337.7522
## 17 3.5 0.1523012 271.3260 83.44846 355.1384
## 18 4.0 0.1500474 278.6769 87.90230 371.9197
## 19 4.5 0.1475359 285.9685 92.38584 388.2582
## 110 5.0 0.1448956 293.2009 96.92876 404.3310
results[which.min(results$roughness), ]
## dist layer temp speed roughness
## 1 0 0.106 221.5 64 170.58
# 2. MÔ HÌNH HỒI QUY DỰ BÁO CHO ĐỘ BỀN VÀ ĐỘ GIÃN (STRENGTH & ELONGATION)
# Tension Strength
model_tension_strenght <- lm(tension_strenght~layer_height+nozzle_temperature+wall_thickness+infill_density+
infill_pattern+bed_temperature+print_speed+material,data=printer_data)
best_model_tension_strenght <- step(model_tension_strenght)
## Start: AIC=179.99
## tension_strenght ~ layer_height + nozzle_temperature + wall_thickness +
## infill_density + infill_pattern + bed_temperature + print_speed +
## material
##
## Df Sum of Sq RSS AIC
## - print_speed 1 8.38 1284.9 178.32
## - infill_pattern 1 15.01 1291.5 178.58
## <none> 1276.5 179.99
## - material 1 128.58 1405.1 182.79
## - bed_temperature 1 139.90 1416.4 183.19
## - nozzle_temperature 1 250.56 1527.1 186.95
## - wall_thickness 1 348.52 1625.0 190.06
## - layer_height 1 588.52 1865.0 196.95
## - infill_density 1 708.24 1984.7 200.06
##
## Step: AIC=178.32
## tension_strenght ~ layer_height + nozzle_temperature + wall_thickness +
## infill_density + infill_pattern + bed_temperature + material
##
## Df Sum of Sq RSS AIC
## - infill_pattern 1 16.38 1301.3 176.95
## <none> 1284.9 178.32
## - material 1 127.94 1412.8 181.07
## - bed_temperature 1 138.99 1423.9 181.46
## - nozzle_temperature 1 249.02 1533.9 185.18
## - wall_thickness 1 491.77 1776.6 192.52
## - layer_height 1 625.74 1910.6 196.16
## - infill_density 1 712.62 1997.5 198.38
##
## Step: AIC=176.95
## tension_strenght ~ layer_height + nozzle_temperature + wall_thickness +
## infill_density + bed_temperature + material
##
## Df Sum of Sq RSS AIC
## <none> 1301.3 176.95
## - material 1 125.70 1427.0 179.56
## - bed_temperature 1 134.29 1435.5 179.86
## - nozzle_temperature 1 242.85 1544.1 183.51
## - wall_thickness 1 476.21 1777.5 190.55
## - layer_height 1 620.46 1921.7 194.45
## - infill_density 1 752.02 2053.3 197.76
summary(best_model_tension_strenght)
##
## Call:
## lm(formula = tension_strenght ~ layer_height + nozzle_temperature +
## wall_thickness + infill_density + bed_temperature + material,
## data = printer_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -9.4717 -4.1695 0.0914 3.8586 13.9581
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 167.02463 53.19880 3.140 0.003055 **
## layer_height 56.36695 12.44845 4.528 4.67e-05 ***
## nozzle_temperature -1.02890 0.36321 -2.833 0.006996 **
## wall_thickness 1.11169 0.28024 3.967 0.000271 ***
## infill_density 0.16643 0.03339 4.985 1.06e-05 ***
## bed_temperature 0.98346 0.46685 2.107 0.041026 *
## materialpla -17.10365 8.39202 -2.038 0.047722 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 5.501 on 43 degrees of freedom
## Multiple R-squared: 0.6667, Adjusted R-squared: 0.6201
## F-statistic: 14.33 on 6 and 43 DF, p-value: 6.783e-09
# Model rút gọn
best_model2 <- lm(tension_strenght~layer_height+nozzle_temperature+wall_thickness+infill_density,
data=printer_data)
summary(best_model2)
##
## Call:
## lm(formula = tension_strenght ~ layer_height + nozzle_temperature +
## wall_thickness + infill_density, data = printer_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -9.5700 -3.8045 -0.5332 3.0743 14.1982
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 61.73141 12.68790 4.865 1.44e-05 ***
## layer_height 56.81320 12.77956 4.446 5.67e-05 ***
## nozzle_temperature -0.27893 0.05672 -4.918 1.21e-05 ***
## wall_thickness 1.16029 0.28621 4.054 0.000197 ***
## infill_density 0.15082 0.03308 4.559 3.93e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 5.649 on 45 degrees of freedom
## Multiple R-squared: 0.6322, Adjusted R-squared: 0.5995
## F-statistic: 19.34 on 4 and 45 DF, p-value: 2.566e-09
# Elongation
model_elongation <- lm(elongation~layer_height+nozzle_temperature+wall_thickness+infill_density+
infill_pattern+bed_temperature+print_speed+material,data=printer_data)
best_model_elongation <- step(model_elongation)
## Start: AIC=-69.64
## elongation ~ layer_height + nozzle_temperature + wall_thickness +
## infill_density + infill_pattern + bed_temperature + print_speed +
## material
##
## Df Sum of Sq RSS AIC
## - infill_pattern 1 0.0446 8.7088 -71.384
## - wall_thickness 1 0.2900 8.9542 -69.995
## <none> 8.6643 -69.641
## - print_speed 1 0.3924 9.0567 -69.426
## - material 1 1.3661 10.0304 -64.320
## - bed_temperature 1 1.5036 10.1678 -63.640
## - infill_density 1 2.6119 11.2762 -58.467
## - nozzle_temperature 1 2.7487 11.4130 -57.864
## - layer_height 1 7.7548 16.4191 -39.679
##
## Step: AIC=-71.38
## elongation ~ layer_height + nozzle_temperature + wall_thickness +
## infill_density + bed_temperature + print_speed + material
##
## Df Sum of Sq RSS AIC
## - wall_thickness 1 0.2624 8.9713 -71.900
## <none> 8.7088 -71.384
## - print_speed 1 0.4092 9.1180 -71.089
## - material 1 1.3548 10.0636 -66.155
## - bed_temperature 1 1.4804 10.1893 -65.534
## - nozzle_temperature 1 2.7193 11.4281 -59.797
## - infill_density 1 2.7397 11.4485 -59.708
## - layer_height 1 7.7218 16.4306 -41.644
##
## Step: AIC=-71.9
## elongation ~ layer_height + nozzle_temperature + infill_density +
## bed_temperature + print_speed + material
##
## Df Sum of Sq RSS AIC
## <none> 8.9713 -71.900
## - print_speed 1 0.9175 9.8887 -69.031
## - material 1 1.4352 10.4065 -66.480
## - bed_temperature 1 1.5981 10.5694 -65.703
## - nozzle_temperature 1 2.9072 11.8785 -59.865
## - infill_density 1 3.0158 11.9871 -59.410
## - layer_height 1 7.4871 16.4584 -43.559
summary(best_model_elongation)
##
## Call:
## lm(formula = elongation ~ layer_height + nozzle_temperature +
## infill_density + bed_temperature + print_speed + material,
## data = printer_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.12952 -0.32192 0.05021 0.30777 0.83809
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 19.023103 4.384019 4.339 8.51e-05 ***
## layer_height 6.079479 1.014849 5.991 3.77e-07 ***
## nozzle_temperature -0.112096 0.030029 -3.733 0.000551 ***
## infill_density 0.010456 0.002750 3.802 0.000448 ***
## bed_temperature 0.106920 0.038632 2.768 0.008295 **
## print_speed -0.004639 0.002212 -2.097 0.041909 *
## materialpla -1.824195 0.695511 -2.623 0.012016 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4568 on 43 degrees of freedom
## Multiple R-squared: 0.7053, Adjusted R-squared: 0.6642
## F-statistic: 17.15 on 6 and 43 DF, p-value: 5.335e-10
# Model rút gọn
best_model3 <- lm(elongation~layer_height+nozzle_temperature+infill_density+print_speed,
data=printer_data)
summary(best_model3)
##
## Call:
## lm(formula = elongation ~ layer_height + nozzle_temperature +
## infill_density + print_speed, data = printer_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.92535 -0.37802 -0.00021 0.34897 1.00524
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.871575 1.058679 7.435 2.32e-09 ***
## layer_height 6.078323 1.076810 5.645 1.05e-06 ***
## nozzle_temperature -0.031633 0.004812 -6.574 4.38e-08 ***
## infill_density 0.008771 0.002824 3.105 0.00328 **
## print_speed -0.004775 0.002346 -2.035 0.04778 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4847 on 45 degrees of freedom
## Multiple R-squared: 0.6528, Adjusted R-squared: 0.6219
## F-statistic: 21.15 on 4 and 45 DF, p-value: 7.235e-10