gopikrsmscs/stock-price-prediction-transformer

Implements a simplified Transformer architecture with multi-head attention and layer normalization for time series forecasting, processing historical OHLCV data normalized via MinMaxScaler into sequential windows. Built on TensorFlow, it handles the complete pipeline from CSV ingestion through sequence generation, model training, and RMSE evaluation with matplotlib visualizations comparing predicted versus actual prices. Adaptable to any stock ticker by swapping the input dataset while maintaining the core Transformer components.

No commits in the last 6 months.