Abstract:
"In the realm of automated news analysis for financial markets, the challenge of efficiently
processing and distilling vast volumes of real-time news data poses significant obstacles for traders and investors. This research endeavours to address this challenge by developing an integrated system that leverages advanced machine learning and natural language processing techniques. The system aims to automate news scraping, summarization, sentiment analysis, and
question-answering, providing timely and insightful information for decision-making in volatile markets.
The methodology employed involves the design and implementation of a tiered architecture
comprising a presentation tier for user interaction, a logic tier housing news scraping and
analysis modules, and a data tier for storing and accessing summarized data. Advanced ML
algorithms are used for text summarization and sentiment analysis, enhancing scalability and accuracy. The system incorporates a chatbot interface for user engagement and further
information retrieval, demonstrating a comprehensive approach to real-time news analysis.
Preliminary results showcase promising performance metrics of the developed system, with
quantitative evaluation using ROUGE scores. The summarization model achieves notable scores, including ROUGE-1 (0.23) and ROUGE-L (0.16), indicating effective compression and retention of key information from news articles. These results validate the efficacy of the system in processing and summarizing real-time news data, setting a foundation for further evaluation and refinement."
