AnalySta: AI-Powered Stock Analysis Predict a...

AnalySta: AI-Powered Stock Analysis Predict and Investment System

Author Name : Dr. M.K. Jayanthi Kannan, Bhavya Singh, Tanvi Sudhir Lalsare, Sanskriti Mittal, Kislay Anand, Moulik Tiwari, Ojas Shukla

DOI: https://doi.org/10.56025/IJARESM.2025.13042502477

ABSTRACT For a rapidly changing and uncertain economic world such as that of financial markets, predicting stock price movements has remained a tough nut to crack. This research proposes an integrated AI-powered model utilizing Long Short-Term Memory (LSTM) networks, sentiment analysis and explainable artificial intelligence (XAI) as a strategy to predict and improve decision-making in stock markets. Historical stock data is used with real-time sentiments derived from news and social media using budding NLP models such as BERT. This is combined with technical indicators and sentiment scores to train LSTM models to produce smart actionable forecasts. SHAP (Shapley Additive Explanations)-to enhance model interpretability so that they can be rationalized as having a basis for decisions such as Buy, Hold, or Sell. The proposed system is made available through an interactive dashboard with real-time visualization and insights. The experimental results show improvement in prediction and model interpretability, thus suggesting the Analysta as a model of prediction power for data-driven investment approaches. ABSTRACT For a rapidly changing and uncertain economic world such as that of financial markets, predicting stock price movements has remained a tough nut to crack. This research proposes an integrated AI-powered model utilizing Long Short-Term Memory (LSTM) networks, sentiment analysis and explainable artificial intelligence (XAI) as a strategy to predict and improve decision-making in stock markets. Historical stock data is used with real-time sentiments derived from news and social media using budding NLP models such as BERT. This is combined with technical indicators and sentiment scores to train LSTM models to produce smart actionable forecasts. SHAP (Shapley Additive Explanations)-to enhance model interpretability so that they can be rationalized as having a basis for decisions such as Buy, Hold, or Sell. The proposed system is made available through an interactive dashboard with real-time visualization and insights. The experimental results show improvement in prediction and model interpretability, thus suggesting the Analysta as a model of prediction power for data-driven investment approaches.