MICE-f: Financial Reviews Analysis using Category Model

In the domain of financial, financial news, articles, reports about financial reviews are helpful and important information which give investors or financial analyst an indication to help decision making in financial matters. However, due to the volume of this information and the diversity of financial topics, it is difficult for a human to track and interpret each of them in a consistent manner. Based on this motivation, we propose that this information can be classified into categories, in which the categories are based on a particular objective and goal as required by a user. In each category, the financial information is further classified based on a status indicator, to reflect the positive, neutral or negative status in term of financial outlook stated in the information. Thus, user can directly focus on interested financial topic, and get an indication about the financial outlook of that topic. For classification purpose, we propose to combine linguistic technique and statistical technique to select features to represent the categories and status indicators. In this work, we present the architecture of MICE-f that will crawl financial information from multiple financial sources and classify them based on the defined categories and status indicators.