##15
```{r classification}
gc()
newData <- dplyr::select(myData,everything()) %>% filter(res_name == 'SO4' | res_name == 'GOL' | res_name == 'EDO' | res_name == 'NAG')
nrow(newData)
newData <- droplevels(newData)
inTrain <- createDataPartition(y = newData$res_name,
                               p = 0.8, list = FALSE)
 training <- newData[inTrain,]
 testing <- newData[-inTrain,]
control <- trainControl(method="cv", number=10)
metric <- "Accuracy"
# a) linear algorithms
#set.seed(100)
fit.lda <- train(res_name ~ res_volume_coverage+local_max_over_std+resolution+local_skewness+local_std+local_mean+local_volume+local_electrons+local_max+res_volume_coverage_second, data=training, method="lda", metric=metric, trControl=control,na.action = na.pass)
# b) nonlinear algorithms
# CART
#set.seed(100)
fit.cart <- train(res_name ~ res_volume_coverage+local_max_over_std+resolution+local_skewness+local_std+local_mean+local_volume+local_electrons+local_max+res_volume_coverage_second, data=training, method="rpart", metric=metric, trControl=control,na.action = na.pass)
# kNN
#set.seed(100)
fit.knn <- train(res_name~res_volume_coverage+local_max_over_std+resolution+local_skewness+local_std+local_mean+local_volume+local_electrons+local_max+res_volume_coverage_second, data=training, method="knn", metric=metric, trControl=control,na.action = na.pass)
# c) advanced algorithms
# SVM
#set.seed(100)
fit.svm <- train(res_name~res_volume_coverage+local_max_over_std+resolution+local_skewness+local_std+local_mean+local_volume+local_electrons+local_max+res_volume_coverage_second, data=training, method="svmRadial", metric=metric, trControl=control,na.action = na.pass)
# Random Forest
#set.seed(100)
fit.rf <- train(res_name~res_volume_coverage+local_max_over_std+resolution+local_skewness+local_std+local_mean+local_volume+local_electrons+local_max+res_volume_coverage_second, data=training, method="rf", metric=metric, trControl=control,na.action = na.pass)

results <- resamples(list(lda=fit.lda, cart=fit.cart, knn=fit.knn, svm=fit.svm, rf=fit.rf))
#results <- resamples(list(lda=fit.lda, cart=fit.cart))
summary(results)

predictions <- predict(fit.rf, testing)
confusionMatrix(predictions, testing$res_name)
```