Stock Price Prediction with LSTM Models

This repository contains a Flask web application for predicting stock prices using Long Short-Term Memory (LSTM) neural networks. The application fetches historical stock data from Yahoo Finance API (yfinance), preprocesses the data by adding moving average features, normalizes it using Min-Max scaling, and trains LSTM models with different hyperparameters.

Key Features

• Data Preprocessing: Includes adding moving average features to enhance model performance.
• Model Training: Utilizes LSTM models with varying architectures (layers, hidden dimensions, dropout) to predict stock prices.
• Hyperparameter Tuning: Grid search over predefined parameters to optimize model performance.
• Web Application: Built using Flask to allow users to input a stock ticker symbol and a date, then receive predictions for future stock prices.
• Visualization: Generates plots using matplotlib to visualize predicted prices and backtesting results.
• Backtesting Strategy: Evaluates a simple trading strategy based on predicted signals.

Technologies Used

• Python
• Flask
• PyTorch (for LSTM models)
• yfinance (Yahoo Finance API)
• matplotlib (for plotting)
• sklearn (for data preprocessing)

How to Use

1. Clone the repository:

   git clone https://github.com/ChellaVigneshKP/stock-prediction.git

2. Start the Flask application:

   python app.py

3. Navigate to http://localhost:5000 in your web browser.
4. Enter a stock ticker symbol and a recent date to get predictions and backtesting results.

Screeshots