A machine learning model that predicts which Tweets are about real disasters and which one’s aren’t. The dataset contains 10,000 classified tweets. I have calculated the frequency of positive(words related to disaster) and negative(words not related to disaster) words in all the tweets, and used a classifier to predict if a tweet is about a real disaster based on these features.
I obtained F1 score of 0.746 on the test data. https://www.kaggle.com/anjalia8/classifyingdisastertweets