AI Intern [Tech Mahindra] Sep-Nov 2024
◦ Data Preprocessing and Feature Engineering : Conducted comprehensive data cleaning by handling missing values and performing feature engineering on dependency ratios to enhance model performance and predictive capabilities.
◦ Model Development and Comparative Analysis: Developed and tuned two advanced predictive models - an LSTM deep learning model and a SARIMAX statistical model - for population trend prediction, utilizing sophisticated interpretation techniques like ShAp and LIME to uncover underlying prediction drivers.
◦ Model Evaluation and Performance: Executed a rigorous evaluation of modeling approaches using performance metrics including RMSE, MAPE, and visual trend analysis, achieving an exceptional 99.63% accuracy in predicting the UK population through the SARIMAX statistical model.