An extraction method of time-series numerical data from press releases by using co-occurrence conditions of numbers and time stamps related to target business keywords

A method of extracting time series data on specific business matters from press releases of a company is proposed. In such press releases, numerical data on specific matters often co‐occur with specific words. From this viewpoint, when keywords are input, the proposed system extracts a set of numerical data and a time stamp that co‐occurs with them. The proposed system extracts numerical data placed close to keywords by using the word distance, and time stamps are extracted in the same manner. The representation of the extracted data is unified for visualization. The results of the application of this method to press releases consisting of several kinds of documents confirm that the accuracy of the extracted numerical data is sufficient to support business analysis. © 2010 Wiley Periodicals, Inc. Electron Comm Jpn, 93(3): 61–69, 2010; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecj.10011