Multivariate Stock Trend & Cryptocurrency Forecasting with Bi-LSTM

Project Overview & Abstract

Advanced time-series forecasting system for AI&DS majors. Employs Bidirectional Long Short-Term Memory (Bi-LSTM) networks paired with VADER NLP sentiment polarity scores scraped from financial news APIs to forecast next-day closing prices with mean absolute error (MAE) under 1.8%.

Technologies & Tools Used

Deliverable Package File Contents