This project proposes a stock price prediction model that integrates financial news data with historical price signals, reframing prediction as a sentiment-based classification task. It utilizes an Average True Range (ATR)-based approach for labeling market sentiment and FinBERT, to enhance alignment with market trends.