Machine learning projects developed for the CodSoft Internship Program. I developed three solutions: Credit card fraud detection, Customer churn prediction, and Movie genre classification.
For the fraud detection and churn prediction projects, I implemented classification techniques and followed a comprehensive ML pipeline that included data preparation, feature selection, and model comparison. I also employed cross-validation with GridSearchCV to fine-tune the models. In the movie genre classification project, I analyzed movie descriptions using natural language processing (NLP) techniques, such as removing stopwords, lemmatizing, and vectorizing significant words.
One major challenge across these projects was class imbalance. I used SMOTE (Synthetic Minority Over-sampling Technique) to balance the datasets, which significantly improved the model performance.