Quantification of Financial News for Economic Surveys

This study concerns financial news articles, which reflect the monetary policy during the US sub prime mortgage crisis. In particular we consider official announcements conducted by the Federal Reserve and its leading representatives. We aim to quantify such information using dependency parsing techniques and statistical measures. In addition, we examine the correlations between the monetary policy and the stock markets by modeling composite index volatilities as functions of key publications. A prototype for the classification of news is targeted, which should reveal the economical impact of events. An eminent aspect of our study is the identification, extraction, and representation of topic-related features and the corresponding instances.