Enhanced Context-Based Stock Recommendation Integrating News Classification and Technical Indicators

This paper presents a method for predicting stock price trends by leveraging online textual news as a contextual factor. Incorporating both historical stock prices and online news data, the cognitive process involves extracting Context Features (CF) from news sources to provide recommendations for traders. Utilizing a Naïve Bayes classification algorithm, news sentiment is efficiently categorized as positive or negative, yielding a News Sentiment Weight (NSW). A News Sentiment Weight (NSW) is then computed for the news, and its impact on a specific stock is assessed to predict the trend. Additionally, the quality of the stock concerning technical is evaluated using different financial technical methods(Indicators), forming the Technical Weight of Stock (TWoS). The combination of NSW and TWoS with weights is utilized to predict the price of the stock trend and offer recommendations to traders. The paper predicts stock trends by combining NSW and TWoS and delivers tailored recommendations to traders. Comparative analysis showcases the superior performance of this contextual approach over traditional recommendation systems, highlighting its efficacy in accurate stock signal prediction.