An Automatic News Analysis and Opinion Sharing System for Exchange Rate Analysis

In an increasingly competitive global market, exchange rate analysis becomes a key part of financial management for international companies. Although some previous studies have developed various forecasting techniques to predict exchange rate trends based on both numerical data and textual data, a deeper understanding of the underlying economic factors is critical for financial decision making purposes. Driven by the real business requirement, we propose an automatic news analysis and opinion sharing system for exchange rate prediction. The system is comprised of five components, including data collection, event classification, sentiment analysis, prediction, and opinion sharing. Breaking news and other data are gathered from broad Web sources in a real time. These news and data are firstly classified into focal event types, an innovative sentiment analysis method is then used to predict their impacts on exchange rate. A built-in mechanism is provided for users to share their opinions. Finally, feedbacks and human judgment from decisions makers are allowed to improve the accuracy. The proposed system has been used in an international logistic company. A real-world case has been used to test and illustrate the effectiveness of the designed system. The paper concludes with a discussion of potential applications.