Naive Bayes Social Network Advertisements

Purpose: Understand how Social media uses data to find new potential client

Step 1: Read Data

# Read csv file
social <- read.csv("Social_Network_Ads.csv")

Step 2 : Prepare data and explore data

summary(social)

##     User.ID            Gender         Age        EstimatedSalary 
##  Min.   :15566689   Female:204   Min.   :18.00   Min.   : 15000  
##  1st Qu.:15626764   Male  :196   1st Qu.:29.75   1st Qu.: 43000  
##  Median :15694342                Median :37.00   Median : 70000  
##  Mean   :15691540                Mean   :37.66   Mean   : 69742  
##  3rd Qu.:15750363                3rd Qu.:46.00   3rd Qu.: 88000  
##  Max.   :15815236                Max.   :60.00   Max.   :150000  
##    Purchased     
##  Min.   :0.0000  
##  1st Qu.:0.0000  
##  Median :0.0000  
##  Mean   :0.3575  
##  3rd Qu.:1.0000  
##  Max.   :1.0000

# User ID column is useless for classfication
social <- social[-1]
social$Purchased <- ifelse(social$Purchased == 0, "No","Yes")
social$Purchased <- factor(social$Purchased, levels = c("No","Yes"))

Step 3 : Split data into training and validation set

library(caTools)
set.seed(2017)
SplitRatio <- 0.75
split = sample.split(social$Purchased, SplitRatio = SplitRatio)
train.sample <- subset(social, split == TRUE)
valid.sample <- subset(social, split == FALSE)

Step 4 : Training a naive bayes Model on training dataset

library(e1071)
classifier <- naiveBayes(Purchased ~ Gender + Age + EstimatedSalary, data = train.sample)

Step 5 : Eveluate the performace of the model

purchased.pred <- predict(classifier, newdata = valid.sample, type = "class")
library(gmodels)
CrossTable( x = valid.sample[,4], 
            y = purchased.pred, 
            prop.chisq = FALSE,
            addmargins = FALSE
)

## 
##  
##    Cell Contents
## |-------------------------|
## |                       N |
## |           N / Row Total |
## |           N / Col Total |
## |         N / Table Total |
## |-------------------------|
## 
##  
## Total Observations in Table:  100 
## 
##  
##                   | purchased.pred 
## valid.sample[, 4] |        No |       Yes | Row Total | 
## ------------------|-----------|-----------|-----------|
##                No |        59 |         5 |        64 | 
##                   |     0.922 |     0.078 |     0.640 | 
##                   |     0.952 |     0.132 |           | 
##                   |     0.590 |     0.050 |           | 
## ------------------|-----------|-----------|-----------|
##               Yes |         3 |        33 |        36 | 
##                   |     0.083 |     0.917 |     0.360 | 
##                   |     0.048 |     0.868 |           | 
##                   |     0.030 |     0.330 |           | 
## ------------------|-----------|-----------|-----------|
##      Column Total |        62 |        38 |       100 | 
##                   |     0.620 |     0.380 |           | 
## ------------------|-----------|-----------|-----------|
## 
##