VECTORVAULT AI-DRIVEN STOCK MARKET PRICE PREDICTOR

Abstract

The Indian stock market, characterized by its non-linear patterns and high volatility across the NSE and BSE, presents a significant challenge for traditional econometric forecasting. This project proposes an advanced, integrated AI-driven framework that transcends conventional statistical limitations by combining quantitative financial data with qualitative sentiment analysis. The system leverages Deep Learning architectures—specifically Long Short-Term Memory (LSTM) and CNN-LSTM hybrids—to model complex temporal dependencies and extract features from historical prices and technical indicators (RSI, MACD, Moving Averages). To capture market psychology, the framework incorporates Natural Language Processing (NLP) using FinBERT and IndicBERT to process multilingual financial news into sentiment scores. Experimental results integrated into a real-time web-based dashboard demonstrate that this multi-modal approach significantly enhances prediction accuracy and provides actionable insights, offering a scalable solution for navigating the complexities of the Indian financial ecosystem.