Predictive_Modeling/Glass_Data_Analysis_1.r at main · Divy0409/Predictive_Modeling · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
# Load required libraries
library(mlbench)  # For dataset
library(ggplot2)  # For plotting
library(corrplot) # For correlation plots
library(e1071)    # For skewness calculation
library(patchwork) # For combining plots
library(caret)    # For data preprocessing

# Load the Glass dataset
data(Glass)

# a) Visualizations and Data Exploration
# Create histograms for each predictor variable
par(mfrow = c(3, 3))  # Arrange the plots in a 3x3 grid

# List of predictor variables
predictors <- names(Glass)[1:9]

# Loop through each predictor and plot the histogram with density line
for (var in predictors) {
  hist_data <- hist(Glass[[var]],
                    main = paste("Histogram of", var),
                    xlab = var,
                    col = "lightblue",
                    border = "black",
                    probability = TRUE)  # Set probability = TRUE for density line

  mean_val <- mean(Glass[[var]], na.rm = TRUE)
  sd_val <- sd(Glass[[var]], na.rm = TRUE)

  curve(dnorm(x, mean = mean_val, sd = sd_val),
        col = "red",
        lwd = 2,
        add = TRUE)
}


# b) Correlation Analysis to find potential outliers and Skewness calculation
# Correlation plot
# Save correlation plot to PDF
pdf("corrplot_glass.pdf", width = 12, height = 10)
corrplot(cor(dplyr::select_if(Glass, is.numeric), use = "na.or.complete"),
         method = 'number',
         type = 'lower',
         diag = FALSE,
         number.cex = 0.75,
         tl.cex = 0.5,
         col = colorRampPalette(c("#00008B", "#0000FF", "#4a1d01", "#FFFF00", "#FF0000", "#8B0000"))(200),  # Darker color palette
         cl.cex = 0.75)  # Adjust the color legend text size
dev.off()

# Boxplots to identify outliers
boxplots <- lapply(predictors, function(var) {
  ggplot(Glass, aes_string(y = var)) +
    geom_boxplot(fill = 'lightblue', color = 'black') +
    labs(title = paste("Boxplot of", var)) +
    theme_minimal()
})

# Arrange all boxplots in a 3x3 grid using patchwork
boxplot_grid <- wrap_plots(boxplots, nrow = 3, ncol = 3)
print(boxplot_grid)

skewness_values <- apply(Glass[, 1:9], 2, skewness)
print("Skewness values for predictor variables:")
print(skewness_values)


# c) Transformation and Outlier Detection

# Apply Box-Cox transformation (excluding the class variable)
trans_boxcox <- preProcess(Glass[, -10], method = c("center", "scale", "BoxCox"))
transformed <- predict(trans_boxcox, Glass[, -10])

# Calculate skewness after transformation
skewness_values_after_box <- apply(transformed, 2, skewness)
print("Skewness after Box-Cox transformation:")
print(skewness_values_after_box)

# Spatial Sign transformation to remove outliers
transformed_spatial <- spatialSign(transformed)

# Skewness after spatial sign transformation
skewness_values_after_spatial <- apply(transformed_spatial, 2, skewness)
print("Skewness after Spatial Sign transformation:")
print(skewness_values_after_spatial)

# Create histograms for each predictor variable after spatial transformation
pdf("Glass_Dataset_Analysis_2.pdf", width = 12, height = 10)
# plot histograms for all the predictor variables
par(mfrow = c(3, 3), mar = c(2, 2, 2, 2))

for (i in 1:9) {
  hist(transformed_spatial[, i],
       main = paste("After Transformation Histogram of", colnames(transformed_spatial)[i]),
       xlab = colnames(transformed_spatial)[i],
       col = "lightblue",
       border = "black",
       prob = TRUE)

  curve(dnorm(x, mean = mean(transformed_spatial[, i], na.rm = TRUE),
              sd = sd(transformed_spatial[, i], na.rm = TRUE)),
        col = "darkblue",
        lwd = 2,
        add = TRUE)
}


dev.off()


# Boxplots to identify outliers after spatial transformation
# Convert transformed_sp to a data frame
transformed_spatial_df <- as.data.frame(transformed_spatial)

# Boxplots to identify outliers after spatial transformation
pdf("boxplots_after_sp.pdf", width = 12, height = 10)

predictors_transformed <- names(transformed_spatial_df)[1:9]

# Create boxplots
boxplots_after_sp <- lapply(predictors_transformed, function(var) {
  # Check for NA values
  if (any(is.na(transformed_spatial_df[[var]]))) {
    message(paste("Warning: NA values found in", var))
  }

  # Create the boxplot
  ggplot(transformed_spatial_df, aes_string(y = var)) +
    geom_boxplot(fill = 'lightblue', color = 'black') +
    labs(title = paste("Boxplot of", var, "(After Spatial Transformation)")) +
    theme_minimal()
})

# Arrange all boxplots in a 3x3 grid using patchwork
boxplot_grid_after_sp <- wrap_plots(boxplots_after_sp, nrow = 3, ncol = 3)
print(boxplot_grid_after_sp)

dev.off()