FinSentix-Sentiment and Term Extraction In Financial Texts

Abstract

This project introduces a web-based platform for financial sentiment evaluation and risk assessment that combines Natural Language Processing (NLP), machine learning techniques, and quantitative finance to generate meaningful insights for investors and financial analysts. The system employs the FinBERT model to examine real-time financial news articles and social media discussions, categorizing stock sentiment into bullish, bearish, or neutral classes. At the same time, it gathers live stock market information and historical price trends through the Yahoo Finance API. A regression-driven risk analysis module calculates key financial indicators, including beta, idiosyncratic risk, market volatility, and the influence of sentiment on price fluctuations. These results are presented through an interactive dashboard that visualizes stock price movements, sentiment trends over the past ten days, and projected market outlooks. By integrating sentiment analysis from NLP, financial data processing, and statistical risk modeling, the platform allows users to understand both behavioral and quantitative factors that influence stock performance. This combined approach helps investors interpret market sentiment alongside numerical indicators, thereby supporting informed investment strategies, effective risk management, and more intelligent financial decision-making