SVM.R

data(cats, package = "MASS")
m <- svm(Sex~., data = cats)
plot(m, cats)


data(iris)
m2 <- svm(Species~., data = iris)
plot(m2, iris, Petal.Width ~ Petal.Length,
slice = list(Sepal.Width = 3, Sepal.Length = 4))
## plot with custom symbols and colors
plot(m, cats, svSymbol = 1, dataSymbol = 2, symbolPalette = rainbow(4),
color.palette = terrain.colors)


data(iris)
attach(iris)
## classification mode
# default with factor response:
model <- svm(Species ~ ., data = iris)
# alternatively the traditional interface:
x <- subset(iris, select = -Species)
y <- Species
model <- svm(x, y, probability = TRUE)
print(model)
summary(model)


#tuning

data(iris)
obj <- tune.svm(Species~., data = iris, sampling = "fix",
gamma = 2^c(-8,-4,0,4), cost = 2^c(-8,-4,-2,0))
plot(obj, transform.x = log2, transform.y = log2)
plot(obj, type = "perspective", theta = 120, phi = 45)


# Importing text data into R package:
# To remove previous objects stored in R environment
rm(list=ls(all=TRUE)) 
#load text mining library
library(tm) 
library(e1071)
#sets R's working directory to near where my files are
#specifies the exact folder where my text file(s) is for analysis with tm.
setwd("C:/Users/gmanish/Desktop/GL_SVM_C4.5/")
docs<-Corpus(DirSource("Docs"), readerControl = list(reader=readPlain)) 
#check what went in
summary(docs)  

#Performing text processing steps:
# Strip any numbers from a text document
docs <- tm_map(docs, removeNumbers)
# Remove punctuation marks from a text document
docs <- tm_map(docs, removePunctuation)
#Strip extra whitespace from a text document
docs <- tm_map(docs , stripWhitespace)
# Remove a set of words from a text document
docs <- tm_map(docs, removeWords, stopwords("english")) 
# this stopword file is at C:\Users\[username]\Documents\R\win-library\2.15.1\library\tm\stopwords 
#Stem words in a text document using Porter's stemming algorithm
docs <- tm_map(docs,stemDocument, language ="english")
#Constructs or coerces to a document-term matrix
dt_matrix <-DocumentTermMatrix(docs) 
# Remove sparse terms from a document-term matrix
dt_matrix <- removeSparseTerms(dt_matrix, 0.75)
#Display detailed information on a document-term matrix
inspect(dt_matrix)


############Data Preparation for Classification Model:
#Store  document-term matrix in a separate data object
matrix=data.frame(inspect(dt_matrix))
data=data.frame(matrix)
#Construct a class variable with levels 1, and 2 for movies, and politics  respectively
data_train=data.frame(rbind(data[1:3,],data[7:9,]))
data_test=data.frame(rbind(data[4:6,],data[10:12,]))
class.train=as.factor(c(1,1,1,2,2,2))
class.test=as.factor(c(1,1,1,2,2,2))

model  =  svm(data_train,class.train, method = "C-classification", kernel = "linear", cost = 10, scale=F)
summary(model)

pred=predict(model,data_test)
table(pred,class.test)
sum(pred==class.test)/length(class.test)

# Model building1 method = "C-classification", kernel = "sigmoid"
model  =  svm(data_train,class.train, method = "C-classification", kernel = "sigmoid", cost = 10, gamma = 0.1)
pred=predict(model,data_test)
table(pred,class.test)
sum(pred==class.test)/length(class.test)

model  =  svm(data_train,class.train, method = "C-classification", kernel = "polynomial", cost = 10, gamma = 0.1)
pred=predict(model,data_test)
table(pred,class.test)
sum(pred==class.test)/length(class.test)