Paper Contents
Abstract
The prediction of stock prices is a challenging task due to the highly volatile and non-linear nature of financial markets. To address this, a web-based Stock Price Prediction System has been developed using deep learning techniques, specifically a recurrent neural network model called Long Short-Term Memory (LSTM). The model is trained on historical closing stock data with a 60-day look-back window to forecast up to 30 days ahead, applying stochastic optimization for weight correction. In addition to prediction, the system integrates key features such as real-time stock retrieval, USD-INR dynamic conversion, market news aggregation, company information access, and a simulated trading environment. The backend uses Flask and MongoDB for secure session management, while the frontend provides an interactive and user-friendly interface with facilities for charting, CSVPDF exports, and transaction history. Experimental results confirm that the model achieves high accuracy in short-term forecasting with mean squared error close to zero, offering improved outcomes compared to traditional approaches. This solution demonstrates the ability of deep learning methods to enhance stock market analysis and decision-making through an integrated, accessible platform.
Copyright
Copyright © 2025 KIRAN M. This is an open access article distributed under the Creative Commons Attribution License.
