Building a text sentiment classifier with the BoW approach

The intent of the BoW approach is to convert the review text provided into a matrix form. It represents documents as a set of distinct words by ignoring the order and meaning of the words. Each row of the matrix represents each review (otherwise called a document in NLP), and the columns represent the universal set of words present in all the reviews. For each document, and across each word, the existence of the word, or the frequency of the word occurrence, in that specific document is recorded. Finally, the matrix created from word frequency vectors represents the documents set. This methodology is used to create input datasets that are required to train the models, and also to prepare the test dataset that need to be used by the trained models to perform text classification. Now that we understand the BoW motivation...