Shuvam Maity

• Developed and compared time series models (Random Forest, Prophet, LSTM) to forecast TMT steel prices of IF and BF routes with an R² of up to 0.98.

• Conducted data preprocessing and feature engineering, such as using lag variables, rolling averages, seasonality, and correlated external variables to optimize the accuracy of the multivariate forecast of stock prices.

• Developed and implemented recursive forecasting for 7, 30, and 60-day periods and examined seasonal and route based price trends.

• Skills and Tools: MS Excel, Python, Numpy, Panda, Scikit-learn, Matplot-lib, Prophet