Part 1 | Disaster Tweets Classification Using Machine Learning with NLP | NLP Project in Python

In this video, we have demonstrated the training of a machine learning model using natural language processing for detecting disaster tweets from the Twitter dataset. Further, we have demonstrated different data pre-processing techniques for cleaning the data. In the end, we have trained two machine learning models i.e., Multinomial Naive Bayes and Passive Aggressive Classifier on Bi-gram and Tr-gram variants of TFIDF vectorized data and found Passive Aggressive Classifier trained on Trigram performed best for this use-case. Lastly, we have also extracted important features of the model for both disaster and normal tweets classes and performed predictions on sample test sentences to check the overall performance of the model.

Blog(Technical write-up): https://wisdomml.in/disaster-tweets-c...

Git Repo (Source Code): https://github.com/wisdomml2020/disas...