rm(list=ls(all=TRUE))
setwd("C:/Users/luigi/Dropbox/TOPIC MODEL")
getwd()
library(quanteda)
library(readtext)
library(caTools)
library(randomForest)
library(caret)
library(naivebayes)
library(car)
library(reshape2)
library(gridExtra)
library(cvTools)

#####################################################
# FIRST STEP: let's prepare the training-set
#####################################################

# let's focus on MOVIE reviews (our DV: either a positive or a negative movie-review)
x <- read.csv("train_review2.csv", stringsAsFactors=FALSE)
str(x)
myCorpusTwitterTrain <- corpus(x)
tok2 <- tokens(myCorpusTwitterTrain, remove_punct = TRUE, remove_numbers=TRUE, remove_symbols = TRUE, split_hyphens = TRUE, 
remove_separators = TRUE)
tok2 <- tokens_remove(tok2, stopwords("en"))
tok2 <- tokens_wordstem (tok2)

Dfm_train <- dfm( tok2)
# Let's trim the dfm in order to keep only tokens that appear in at least 5% of the reviews
Dfm_train <- dfm_trim(Dfm_train , min_docfreq = 0.05, verbose=TRUE, docfreq_type = "prop")
topfeatures(Dfm_train , 20)  # 20 top words

# Always check if after trimming your test-set you have some texts with just 0s! 
Dfm_train [ntoken(Dfm_train ) == 0,]

# our classes
table(Dfm_train@docvars$Sentiment)
# our benchmark: accuracy .524
prop.table(table(Dfm_train@docvars$Sentiment))

train <- as.matrix(Dfm_train) 

######################################################
######################################################
# Let's make CROSS-VALIDATION with k-fold=5 with a loop!
# HERE you HAVE JUST 2 classes for the outcome (negative/positive)
######################################################
######################################################

# STEP 1: create the 5 folds
ttrain <- train # let's change the name of the original train matrix, given that we are already going to use such name below in the loop
# let's split our training-set in 5 folds 
set.seed(123) # set the see for replicability
k <- 5 # the number of folds; it does not matter the number of folds you decide here; the below procedure always will work!
folds <- cvFolds(NROW(ttrain ), K=k)
str(folds)

######################################################
######################################################
# Let's start with a Random Forest loop
######################################################
######################################################

# STEP 2: the LOOP

# here we fix ntree=100 to make things faster!

system.time(for(i in 1:k){
  train <- ttrain [folds$subsets[folds$which != i], ] 
  validation <- ttrain [folds$subsets[folds$which == i], ]
  set.seed(123)
  newrf <- randomForest(y= as.factor(Dfm_train[folds$subsets[folds$which != i], ]@docvars$Sentiment) ,x=train, do.trace=TRUE, ntree=100) 
  set.seed(123)
  newpred <- predict(newrf,newdata=validation) 
 class_table <- table("Predictions"= newpred, "Actual"=Dfm_train[folds$subsets[folds$which == i], ]@docvars$Sentiment)
 print(class_table)  
  df<-confusionMatrix( class_table, mode = "everything") 
  df.name<-paste0("conf.mat.rf",i)
  assign(df.name,df)
})

# STEP 3: the metrics
ls()
# we have created 5 objects that have saved the 5 confusion matrices we have created. 
# We can estimate now the performance metrics on such results. For example:
conf.mat.rf1
str(conf.mat.rf1)
conf.mat.rf1$overall[1] # overall accuracy: (32+49)/(32+49+13+6)
# note that the F1 value you see is not the one for the overall model, but just for the first class (i.e., "negative")
# Therefore we have to estimate the average value of F1 for each k-fold by hands! See below

conf.mat.rf1$byClass[1] # Recall for negative: (32)/(32+13) - think vertically!
conf.mat.rf1$byClass[3] # Precision for negative: (32)/(32+6) - think horizontally!
conf.mat.rf1$byClass[2] # Recall for positive: (49)/(49+6) - think vertically!
conf.mat.rf1$byClass[4] # Precision for positive: (49)/(49+13) - think horizontally!

# F1 per the negative class: 
(2*conf.mat.rf1$byClass[3]*conf.mat.rf1$byClass[1])/(conf.mat.rf1$byClass[3]+conf.mat.rf1$byClass[1]) # F1 for negative
# and for positive?
(2*conf.mat.rf1$byClass[2]*conf.mat.rf1$byClass[4])/(conf.mat.rf1$byClass[2]+conf.mat.rf1$byClass[4]) # F1 for positive

RFPredict <- data.frame(col1=vector(), col2=vector(), col3=vector()) 

for(i in  mget(ls(pattern = "conf.mat.rf")) ) {
Accuracy <-(i)$overall[1] # save in the matrix the accuracy value
F1_negative <- (2*(i)$byClass[1]*(i)$byClass[3])/((i)$byClass[1]+(i)$byClass[3]) # save in the matrix the F1 value for negative
F1_positive <- (2*(i)$byClass[2]*(i)$byClass[4])/((i)$byClass[2]+(i)$byClass[4]) # save in the matrix the F1 value for positive
RFPredict <- rbind(RFPredict , cbind(Accuracy , F1_negative, F1_positive))
 }

str(RFPredict )

# Let's compute the average value for accuracy and f1

acc_rf_avg <- mean(RFPredict [, 1] )
f1_rf_avg <- mean(colMeans(RFPredict [-1] ))

acc_rf_avg
f1_rf_avg

# the fact that the avg. accuracy and the avg. F1 values are close to each other is of course a good news!

######################################################
######################################################
# Naive Bayes LOOP
######################################################
######################################################

system.time(for(i in 1:k){
  train <- ttrain [folds$subsets[folds$which != i], ] 
  validation <- ttrain [folds$subsets[folds$which == i], ]
  newrf <-  multinomial_naive_bayes(y= as.factor(Dfm_train[folds$subsets[folds$which != i], ]@docvars$Sentiment) ,x=train,  laplace = 1) 
  newpred <- predict(newrf,newdata=validation, type="class") 
class_table <- table("Predictions"= newpred, "Actual"=Dfm_train[folds$subsets[folds$which == i], ]@docvars$Sentiment)
print(class_table)  
  df<-confusionMatrix( class_table, mode = "everything") 
  df.name<-paste0("conf.mat.nb",i)
  assign(df.name,df)
})

# STEP 3: the metrics
NBPredict <- data.frame(col1=vector(), col2=vector(), col3=vector()) 

for(i in  mget(ls(pattern = "conf.mat.nb")) ) {
Accuracy <-(i)$overall[1] # save in the matrix the accuracy value
F1_negative <- (2*(i)$byClass[1]*(i)$byClass[3])/((i)$byClass[1]+(i)$byClass[3]) # save in the matrix the F1 value for negative
F1_positive <- (2*(i)$byClass[2]*(i)$byClass[4])/((i)$byClass[2]+(i)$byClass[4]) # save in the matrix the F1 value for positive
NBPredict <- rbind(NBPredict , cbind(Accuracy , F1_negative, F1_positive))
 }

str(NBPredict )

# Let's compute the average value for accuracy and f1

acc_nb_avg <- mean(NBPredict [, 1] )
f1_nb_avg <- mean(colMeans(NBPredict [-1] ))

acc_nb_avg
f1_nb_avg 

######################################################
######################################################
# Let's compare the results we got via Naive Bayes, SVM and RF 
######################################################
######################################################

acc_rf_avg
acc_nb_avg

f1_rf_avg
f1_nb_avg

# Let's plot the results!

gb1 <- as.data.frame(acc_nb_avg )
colnames(gb1)[1] <- "Accuracy NB"
gb2 <- as.data.frame(acc_rf_avg )
colnames(gb2)[1] <- "Accuracy RF"
ac_tot <- cbind(gb1, gb2)
ac_tot
str(ac_tot)
df.long_ac_tot<-melt(ac_tot)
str(df.long_ac_tot)

p <- ggplot(df.long_ac_tot, aes(x=variable, y=value, color = variable)) +
geom_boxplot() + xlab("Algorithm") + ylab(label="Values of Accuracy") +
labs(title = "Cross-validation with k =5: values of Accuracy") + coord_flip() +
theme_bw() + theme(legend.position = "None")

gb1 <- as.data.frame(f1_nb_avg)
colnames(gb1)[1] <- "F1 NB"
gb2 <- as.data.frame(f1_rf_avg )
colnames(gb2)[1] <- "F1 RF"
f1_tot <- cbind(gb1, gb2)
f1_tot
str(f1_tot)
df.long_f1_tot<-melt(f1_tot)
str(df.long_f1_tot)

p2 <- ggplot(df.long_f1_tot, aes(x=variable, y=value, color = variable)) +
geom_boxplot() + xlab("Algorithm") + ylab(label="Values of F1") +
labs(title = "Cross-validation with k =5: values of F1") + coord_flip() +
theme_bw() + theme(legend.position = "None")

grid.arrange(p, p2,  nrow=2) # Plot everything together

# So in this particular case, NB appears to be better than the RF! 
# But remember: you are employing a training-set consisting of just 500 documents 
# and NB usually does well in this scenario. Moreover...
# Wait a minute! In this example we kept the tuning parameters fixed! 
# But is there any other tuning parameters configuration for RF and NB that will improve the model?