ToxicCommentClassification

This project is an implementation of toxic comment classification challenge in Kaggle and it is hosted here as a consumable application.

This project has three parts

1. Preprocessing

2. Modeling

3. Dash Application

Preprocessing:

Removed stopwords, punctuations, blank lines and some urls, hyperlinks and IPs from the input texts. Used WordNetLemmatizer to lemmatize the words and used glove 100d word vectors as embeddings.

Models:

Built three different models using Keras. It includes a CNN, a RNN and a Naive Bayes SVM. These model outputs are stacked to get the final output.

Dash Application:

A consumable UI is created using dash and is hosted in heroku.