Generating Concept Hierarchies from Text for Intelligence Analysis

It is important to automatically extract key information from sensitive text documents for intelligence analysis. Text documents are usually unstructured and information extraction from texts usually focuses on named entity extraction that is insufficient to reflect the concepts contained within documents. Other research on texts focuses on classifying the whole text into categories. However, texts might be diverse in subjects and suffer from classification accuracy. In this paper, our goal is to present a feasible approach to extract relevant information from texts and organize into concept hierarchies with the help of Web corpora. The approach is designed considering the need of information intelligence and the adaptation to different application domains.

• Type: Conference
• Year: 2004
• Where: ISI
• Authors: Jenq-Haur Wang, Chien-Chung Huang, Jei-Wen Teng, Lee-Feng Chien
• Keywords: ISI 2004, Sensitive Text Documents, Text Documents, Texts