A Novel CNN-SVM Framework for Stock Trend Prediction Leveraging Historical Market Data and News Sentiment

The stock market direction requires advanced expertise due to numerous elements including market price movement histories along with financial news displays. The authors introduce an AI model that unites deep learning approaches with machine learning methods for improved stock market trend analysis. The proposed model applies Support Vector Machines (SVMs) to make trend classifications after Convolutional Neural Networks (CNNs) extract necessary stock market data features. The model analyzed financial news sentiment to determine public sentiment effects on stock price shifts by integrating numerical stock data with text data during the prediction process. The CNN-SVM model succeeded in reaching a prediction accuracy rate of 89.2% while surpassing standalone CNN at 84.5% in addition to Random Forest at $\mathbf{8 1. 7 \%}$ and ARIMA at $\mathbf{7 8. 3 \%}$. Market trends get primarily shaped by closing price movements and sentiment scores according to the analysis results. The stock prices of positively rated companies experienced a 2.8% increase during the following trading day whereas negatively assessed stocks decreased by 3.4%. Testing showed the model maintained a correct relationship between market sentiment and stock prices even though it produced some minor labeling errors which led to enhanced predictive ability. Artificial intelligence hybrid forecasting models have proven their exceptional capability in financial market predictions. The approach which combines historical price data with financial news sentiment offers investors useful findings that help decrease uncertainty while providing evidence-based investment choices. Deep learning methodologies integrated with established ML techniques produce efficient market prediction models which constitute a usable tool for traders and financial analysts.