Textometry and Information Discovery: A New Approach to Mining Textual Data on the Web

摘要

Most Text Mining tasks focus on local linguistic rules for detecting such elements as named entities, events and opinions: the goal here is to go beyond these local context boundaries by taking global dimensions into account. A robust method to mine textual data known as Textometry is not constrained by external resources and avoids problems such as the coverage limitations of standard dictionaries and at a higher level, domain-dependant resources. Textometry provides a new approach of exploring and comparing textual data. This paper studies the Textometric method and how it can be applied to the industrial context of mining named entities and their trends (opinions or events) in both French and American online news media: Le Monde and the New York Times. This paper focuses on bypassing certain costly steps in tasks related to mining information on Named Entities.