Student Performance Predictor Application

• End-to-end application that predicts the student performance based on his/her prior educational and geographic records. The dataset was created synthetically for training and analyzed using seaborn library. • The application was built using Python and trained using a Random Forest model and...

Recommender System for an E-Commerce Application

• A recommender system with applied Ensemble Learning, featuring matrix factorization clubbed with bagging for optimized result • Addressed common challenges for recommender systems such as Cold Start problem, Data sparsity, accuracy, scalability and diversity, time complexity & compute intensity

Prediction of Incidence of Alzheimer’s Disease(AD) using Neural Networks

• Predicting the incidence of AD given his/her health and socio-economic history • Modeled a neural network based on generated polynomial features and strategized optimization to overcome high bias and other accuracy concerns, resulting in 82.16% of CV accuracy

Music Recommendation System Based on Current Mood

• Incorporated Content Based Filtering in which the system modelled Clustering using K-Means, Data Pre-processing, Feature extraction, Mood identification using feature analysis and Song recommendation