THMMY.gr

 model_svm=svm(Species~.,data=trainingdata,kernel="radial",type="C-classification",gama=0.1)
pred=predict(model_svm,testingdata[,c(1,2)])
Accuracy(pred,testingdata$Species)

> X1=4;
> X2=5;
> X3=6;
> tryval=expand.grid(X1,X2,X3)
> colnames(tryval)=colnames(iris2)
> tryval
  Sepal.Length Sepal.Width Species
1            4           5       6

> test=predict(model_svm,tryval[,c("Sepal.Length","Sepal.Width")])
> test
     1 
setosa 
Levels: setosa versicolor

TP=sum((P_M1>0.5 & Class>0)==TRUE)
FN=sum((P_M1<0.5 &Class==1)==TRUE)

pca_model <- prcomp(training, center = TRUE, scale = TRUE)
eigenvectors = pca_model$rotation
eigenvalues = pca_model$sdev^2
acc = c()
for (i in 1:9){
  training_pc <- as.data.frame(predict(pca_model, training)[, 1:i])
  training_pc[, i+1:9] <- 0
  
  testing_pc <- as.data.frame(predict(pca_model, testing)[, 1:i])
  testing_pc[, i+1:9] <- 0
  
  knn_model <- knn(train = training_pc, test = testing_pc, cl = trainingType, k = 3, prob = TRUE)
  acc[i] = Accuracy(y_pred = knn_model, y_true = testingType)
}

pca_model <- prcomp(training, center = TRUE, scale = TRUE)
eigenvectors = pca_model$rotation
eigenvalues = pca_model$sdev^2
acc = c()
for (i in 1:9){
  training_pc <- as.data.frame(predict(pca_model, training)[, 1:i])
  training_pc[, i+1:9] <- 0
  
  testing_pc <- as.data.frame(predict(pca_model, testing)[, 1:i])
  testing_pc[, i+1:9] <- 0
  
  knn_model <- knn(train = training_pc, test = testing_pc, cl = trainingType, k = 3, prob = TRUE)
  acc[i] = Accuracy(y_pred = knn_model, y_true = testingType)
}

testing_pc[, i+1:9] <- 0

training_pc[, i+1:9] <- 0

testing_pc[, (i+1):9] <- 0

Accuracy(cluster, target)

cluster = cutree(hc_model, k = 2)

library(MLmetrics)

data(Glass, package = "mlbench")
training = Glass[c(1:50, 91:146), -10] 
trainingType = factor(Glass[c(1:50, 91:146), 10]) 
testing = Glass[51:90, -10] 
testingType = factor(Glass[51:90, 10])

pca_model = prcomp(training, center=TRUE, scale=TRUE)
eigenvals = pca_model$sdev^2
eigenvecs = pca_model$rotation
ans1 = eigenvals[1] / sum(eigenvals)

ans2 = 1 - sum(eigenvals[1:4])/sum(eigenvals)

pred = knn(training, testing, trainingType, k=3)
ans3 = Accuracy(pred, testingType)

ans4 = Recall(testingType, pred, positive=2)

acc = c()
for(comp in 1:9){
  reduced_tr = as.data.frame(predict(pca_model, training)[, 1:comp])
  reduced_tst = as.data.frame(predict(pca_model, testing)[, 1:comp])
  
  pred = knn(reduced_tr, reduced_tst, trainingType, k=3)
  acc[comp] = Accuracy(pred, testingType)
  
}