Trading Price Prediction Using Sentimental Analysis

DOI

https://www.doi.org/10.59256/indjcst.20260502072

Published In

Volume 5, Issue 2 (2026)

Published Online

May-August 2026

Pages

628-632

Article Type

Original Article

Check for updates

The work demonstrates a new method of a commercial price predictor that unites both technical analysis and ML with asset research on the markets of equities, foreign exchange, and cryptocurrency. This approach is used to determine market trends and volatility by using historical price data along with technical indicators such as moving averages, Bollinger Bands, the MaCD, and the RSI. To evaluate the predictive power of the different ML methods, such metrics as score r 2, MAE, and MAPE are utilized to evaluate the prediction of various ML methods, which are linear regression, random forest, XGBoost, and KNN. The inclusion of sentiment research in Finber and rules-based categorization is an important modification that evaluates the effect of financial reporting on the dynamics of markets in real-time. Emotional module calculates the general sentiment score of the market. This system has been developed as a simplified web application that offers interactive visualization, performance comparison between different models, and adaptation to the data files through uploaded users and data stream in real-time. In the case of traders who track the overall knowledge based on the combined analysis of technical and emotions, modular architecture ensures the ability to expand and be useful in practice.