SVM-Iris-Dataset.R

#set WD
setwd("C:/Users/sharl/Desktop/USF/Fall 2021/LIS 4805 - Predictive Analytics/Week 10 - SVM")

#support vector machine using iris data
library(e1071)
plot(iris)

plot(iris$Sepal.Length, iris$Sepal.Width, col=iris$Species)

plot(iris$Petal.Length, iris$Petal.Width, col=iris$Species)

length(iris)

## [1] 5

dim(iris)

## [1] 150   5

s <-sample(150, 100)
col<-c("Petal.Length", "Petal.Width", "Species")
iris_train <-iris[s,col]
iris_test <-iris[-s,col]

svmfit <-svm(Species ~., data=iris_train, kernel="linear", cost=0.1, scale=F)
print(svmfit)

## 
## Call:
## svm(formula = Species ~ ., data = iris_train, kernel = "linear", 
##     cost = 0.1, scale = F)
## 
## 
## Parameters:
##    SVM-Type:  C-classification 
##  SVM-Kernel:  linear 
##        cost:  0.1 
## 
## Number of Support Vectors:  46

plot(svmfit, iris_train[,col])

#cross-validation
tuned <-tune(svm, Species~., data=iris_train, kernel="linear", ranges=list(cost=c(0.001, 0.01, 0.1, 1,10,100)))
summary(tuned)

## 
## Parameter tuning of 'svm':
## 
## - sampling method: 10-fold cross validation 
## 
## - best parameters:
##  cost
##   0.1
## 
## - best performance: 0.05 
## 
## - Detailed performance results:
##    cost error dispersion
## 1 1e-03  0.78 0.06324555
## 2 1e-02  0.37 0.11595018
## 3 1e-01  0.05 0.05270463
## 4 1e+00  0.05 0.05270463
## 5 1e+01  0.05 0.05270463
## 6 1e+02  0.06 0.05163978

#set cost=100
svmfit <-svm(Species ~., data=iris_train, kernel="linear", cost=100, scale=F)
print(svmfit)

## 
## Call:
## svm(formula = Species ~ ., data = iris_train, kernel = "linear", 
##     cost = 100, scale = F)
## 
## 
## Parameters:
##    SVM-Type:  C-classification 
##  SVM-Kernel:  linear 
##        cost:  100 
## 
## Number of Support Vectors:  11

plot(svmfit, iris_train[,col])

p <-predict(svmfit, iris_test[,col], type="class")
plot(p)

table(p, iris_test[,3])

##             
## p            setosa versicolor virginica
##   setosa         15          0         0
##   versicolor      0         16         2
##   virginica       0          0        17

mean(p==iris_test[,3])

## [1] 0.96