Model7.r

setwd("/media/shilpi/New Volume/ML/r/breast_cancer");
data=read.csv("breastcancer.csv")
data
str(data)
data.names();
summary(data)


#PRE PROCESSING
data$Clump_Thickness=as.numeric(data$Clump_Thickness)
data$Uniformity_CellSize=as.numeric(data$Uniformity_CellSize)
data$Uniformity_CellShape=as.numeric(data$Uniformity_CellShape)
data$Marginal_Adhesion=as.numeric(data$Marginal_Adhesion)
data$Single_Epithelial_cellSize=as.numeric(data$Single_Epithelial_cellSize)
data$BareNuclei=as.numeric(data$BareNuclei)
data$Bland_Chromatin=as.numeric(data$Bland_Chromatin)
data$Normal_Nucleoli=as.numeric(data$Normal_Nucleoli)
data$Mitoses=as.numeric(data$Mitoses)
data$Class=as.numeric(data$Class)


data[is.na(data$Clump_Thickness)]<-mean(data$Clump_Thickness,na.rm=TRUE)
data[is.na(data$Uniformity_CellSize)]<-mean(data$Uniformity_CellSize,na.rm = TRUE)
data[is.na(data$Uniformity_CellShape)]<-mean(data$Uniformity_CellShape,na.rm = TRUE)
data[is.na(data$Marginal_Adhesion)]<-mean(data$Marginal_Adhesion,na.rm = TRUE)
data[is.na(data$Single_Epithelial_cellSize)]<-mean(data$Single_Epithelial_cellSize,na.rm = TRUE)
data[is.na(data$BareNuclei)]<-mean(data$BareNuclei,na.rm = TRUE)
data[is.na(data$Bland_Chromatin)]<-mean(data$Bland_Chromatin,na.rm = TRUE)
data[is.na(data$Normal_Nucleoli)]<-mean(data$Normal_Nucleoli,na.rm = TRUE)
data[is.na(data$Mitoses)]<-mean(data$Mitoses,na.rm = TRUE)
data[is.na(data$Class)]<-2
summary(data)

set.seed(100)



sub <- sample(nrow(data), floor(nrow(data) * 0.75))
training <- data[sub,(3:12) ]
testing <- data[-sub, (3:11)]

real<- data[-sub,12]
summary(training)

library(mlbench)
library("Boruta")
typeof(data)
boruta.train <- Boruta(data$Class ~ ., data=data ,doTrace = 2,ntree=5)

boruta.train

#removing 7,mitoses
data$Single_Epithelial_cellSize<-NULL
data$Mitoses<-NULL
summary(data)

malignantdata=subset(data,Class==2)
benigndata=subset(data,Class==4)
str(malignantdata)

k=2;
library("fclust")
KMC = FKM(malignantdata[ ,3:11], k= 2,m=2)
KMC$clus[ ,1]




malignantdata$Class=KMC$clus[ ,1]
benigndata$Class=3


data<-rbind(malignantdata,benigndata)
write.csv(data, file = "BreastCancer4.csv",row.names = TRUE)
#Run the fuzzy knn python code on BreastCancer4.csv