Natural Language Processing With Disaster Tweets

In this competition, you’re challenged to build a machine learning model that predicts which Tweets are about real disasters and which one’s aren’t. I had access to a dataset of 10,000 tweets that were hand classified. In order to solve the problem I have build different classification Models.

Telecom Customer Churn Prediction

In this project I have build different ML classification models to be able to predict whether a particular customer is gonna churn or not, I have also tuned these models and used Model explainability techniques like Permutation Importance and SHAP to find the most important features of the dataset.

Fraud Detection in Electricity and Gas Consumption

Using the client’s billing history, the aim of the challenge is to detect and recognize clients involved in fraudulent activities. I have used different Boosting Algorithms out of which LGBM Classifier has outperformed other Models. The most powerful part of this project was the feature engineering.