Decision_Tree

Author

The Null wranglers

Load Libraries 
suppressWarnings(library(tidyverse))
library(knitr)
library(lubridate)
library(ggplot2)
library(dplyr)
library(rpart)
library(rpart.plot)
Load and Filter Data 
data <- read.csv("data/study_data.csv")


# Filter the data to include only classes with one section
filtered_data1 <- data %>% filter(Sections <= 1)


# Categorize DEW_COUNT based on the mean value of Percent.DEW

filtered_data1$DEW_COUNT[filtered_data1$Percent.DEW <= 12.2] <- 'Low'

filtered_data1$DEW_COUNT[filtered_data1$Percent.DEW > 12.2] <- 'High'

# Convert DEW_COUNT to a factor variable
filtered_data1$DEW_COUNT <- as.factor(filtered_data1$DEW_COUNT)

# Select relevant columns for the decision tree
tree_data <- select(filtered_data1, DEW_COUNT, Full_Online, Hybrid, Live_Online, Reg_Session, Monday,Tuesday, Wednesday, Thursday, Friday, Saturday,              Sunday, Early_Morning, Mid_Morning, Early_Afternoon, Mid_Afternoon, Evening, Asynchronous  )

# Convert selected columns to logical (TRUE/FALSE) values
col_names <- c("Full_Online", "Hybrid", "Live_Online", "Reg_Session", "Monday",         "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday",              "Sunday", "Early_Morning", "Mid_Morning", "Early_Afternoon", "Mid_Afternoon", "Evening", "Asynchronous" )
tree_data[col_names] <- sapply(tree_data[col_names], as.logical)
Confirm that NA Values are removed 
# Check for missing values in the dataset
missing_values <- colSums(is.na(tree_data))
print(missing_values)
      DEW_COUNT     Full_Online          Hybrid     Live_Online     Reg_Session 
              0               0               0               0               0 
         Monday         Tuesday       Wednesday        Thursday          Friday 
              0               0               0               0               0 
       Saturday          Sunday   Early_Morning     Mid_Morning Early_Afternoon 
              0               0               0               0               0 
  Mid_Afternoon         Evening    Asynchronous 
              0               0               0
Summarize Data 
# Display a summary of the dataset
summary(tree_data)
 DEW_COUNT   Full_Online       Hybrid        Live_Online     Reg_Session    
 High:1525   Mode :logical   Mode :logical   Mode :logical   Mode :logical  
 Low :2403   FALSE:3126      FALSE:3479      FALSE:2948      FALSE:478      
             TRUE :802       TRUE :449       TRUE :980       TRUE :3450     
   Monday         Tuesday        Wednesday        Thursday      
 Mode :logical   Mode :logical   Mode :logical   Mode :logical  
 FALSE:2716      FALSE:2278      FALSE:2668      FALSE:2285     
 TRUE :1212      TRUE :1650      TRUE :1260      TRUE :1643     
   Friday         Saturday         Sunday        Early_Morning  
 Mode :logical   Mode :logical   Mode :logical   Mode :logical  
 FALSE:3314      FALSE:3928      FALSE:3928      FALSE:3333     
 TRUE :614                                       TRUE :595      
 Mid_Morning     Early_Afternoon Mid_Afternoon    Evening       
 Mode :logical   Mode :logical   Mode :logical   Mode :logical  
 FALSE:3201      FALSE:3316      FALSE:3025      FALSE:3665     
 TRUE :727       TRUE :612       TRUE :903       TRUE :263      
 Asynchronous   
 Mode :logical  
 FALSE:3100     
 TRUE :828
Create and Visualize Tree 
# Build a decision tree
tree_default <- tree_data %>%
  rpart(DEW_COUNT ~ ., data = .)

# Plot the decision tree and save it as a PNG file
png("images/decision_tree.png", width = 1000, height = 600)
rpart.plot(tree_default, box.palette = "auto", shadow.col = "gray", nn = TRUE, roundint = FALSE, cex = 1.5)
dev.off()
png 
  2
Create and Visualize Tree 
# Plot the decision tree
rpart.plot(tree_default, box.palette = "auto", shadow.col = "gray", nn = TRUE, roundint = FALSE, cex = 1.5)

  1. The decision tree predicts “High” if Full_Online is >= 0.5; otherwise, it predicts “Low.”
  2. Tree building stops when additional splits don’t significantly improve performance or when there are insufficient observations for further splitting.