Python for NLP: Creating Multi-Data-Type Classification Models with Keras
This is the 18th article in my series of articles on Python for NLP. In my previous article, I explained how to create a deep learning-based movie sentiment analysis model using Python’s Keras library. In that article, we saw how we can perform sentiment analysis of user reviews regarding different movies on IMDB. We used the text of the review the review to predict the sentiment.
However, in text classification tasks, we can also make use of the non-textual information to classify the text. For instance, gender may have an impact on the sentiment of the review. Furthermore, nationalities may affect the public opinion about a particular movie. Therefore, this associated info, also known as meta data can also be used to improve accuracy of statistical model.
In this article, we will build upon the concepts that we studied in the last two articles and will see how to create a text classification system that classifies user reviews regarding different business, into one of the three predefined categories i.e. “good”, “bad”, and “average”. However, in addition to the text of the review, we will use the associated meta data of the review to perform