Portfolio

In this project I performed Data Mining on Altair AI Studio for three different datasets i,e Tool Forces Prediction model, Tool Wear Detection Time Series model followed by Classification, Machine Usage Condition Data Classification model.

Built a multi-class classification model for early Alzheimer’s stage detection and a temporal forecasting model to predict the evolution of clinical metrics based on patient history.

Conducted a comparative study of traffic forecasting using LSTM and a spatio-temporal Diffusion Convolutional Neural Network on the PEMS-BAY dataset, incorporating a spatially aware adjacency matrix to model traffic flow with diffusion dynamics.

Implemented collaborative filtering techniques, leveraging user-item interaction data to recommend movies and shows based on user preferences. Deployed the recommendation engine as a REST API using Flask and Docker, ensuring scalability and seamless integration into existing systems, boosting user engagement.

The goal of the project was to create and analyze a Transactional Database - IEEE Credit Card fraud detection database. Detection of fraudulent patterns using SQL Analytics. Optimization of Schema Design with partitioning/indexing to handle massive datasets. PL/pgSQL functions for in-database scoring and addressing real-time detection needs.

Led a team of 3 to develop an explainable Predicitve ML model for EV sales covering Spatial analysis of State-Level Infrastructure Density Mapping. Feature Engineering and EDA for all the features followed by Statistical Testing. Evaluated all the models based on the metrics. Also Hosted a Visualization Dashboard on PowerBI.