RohithKoripelli/indian-bank-stock-models

Indian Bank Stock Price Prediction Models

This repository contains 8 trained V5 Transformer models for predicting next-day price movements of major Indian banking stocks.

Models

• HDFC Bank (HDFCBANK.NS)
• ICICI Bank (ICICIBANK.NS)
• Kotak Mahindra Bank (KOTAKBANK.NS)
• Axis Bank (AXISBANK.NS)
• State Bank of India (SBIN.NS)
• Punjab National Bank (PNB.NS)
• Bank of Baroda (BANKBARODA.NS)
• Canara Bank (CANBK.NS)

Model Architecture

• Type: V5 Transformer
• Features: 35 (technical, sentiment, fundamental, macro, sector)
• Lookback: 60 days
• Parameters: ~154,808 per model

Performance

Average metrics across all 8 models:

• MAPE: 0.84%
• R²: 0.9771
• Directional Accuracy: 65.15%

Usage

from huggingface_hub import hf_hub_download
import tensorflow as tf

# Download a specific model
model_path = hf_hub_download(
    repo_id="Rohithkoripelli/indian-bank-stock-models",
    filename="HDFCBANK/best_model.keras"
)

# Load the model
model = tf.keras.models.load_model(model_path)

Training Data

• Date Range: January 2019 - January 2026
• Records: ~1,743 per stock
• Features: Technical indicators, sentiment scores, fundamental metrics

Automation

These models are used in an automated GitHub Actions workflow that:

1. Collects latest stock data daily
2. Downloads models from Hugging Face
3. Generates predictions
4. Deploys to Vercel

License

MIT License - Free to use for research and educational purposes.

Repository

Full code and documentation: Stock-Price-Prediction