Application using text2vec examples

In this section, we will analyze the performance of logistic regression on various examples of text2vec.

How to do it...

Here is how we apply text2vec:

1. Load the required packages and dataset:

library(text2vec) 
library(glmnet) 
data("movie_review") 

2. Function to perform Lasso logistic regression, and return the train and test AUC values:

logistic_model <- function(Xtrain,Ytrain,Xtest,Ytest)
{ 
  classifier <- cv.glmnet(x=Xtrain, y=Ytrain, 
  family="binomial", alpha=1, type.measure = "auc", 
  nfolds = 5, maxit = 1000) 
  plot(classifier) 
  vocab_test_pred <- predict(classifier, Xtest, type = "response") 
  return(cat("Train AUC : ", round(max(classifier$cvm), 4), 
  "Test AUC : ",glmnet:::auc(Ytest, vocab_test_pred),"\n")) 
} 

3. Split the movies review data into train and test in an 80:20 ratio:

train_samples <- caret::createDataPartition(c(1:length(labels[1,1])),p = 0.8)$Resample1 
train_movie <- movie_review[train_samples,] 
test_movie <- movie_review...