A Domain-Oriented Double Dictionary Forward Traversal Method for Fine Corpus Extraction

Abstract

Improving audit intelligence requires computers to understand the semantics of audit information. At present, the researches on the related field of the intelligent information processing show that, the basis of intelligent information processing is the natural language understanding. This paper which combines the construction technology of corpus researches the basic methods and techniques on the construction of audit corpus. According to the text features in social security audit field, the paper proposed dual dictionary secondary forward traversal keyword extraction method which combines the specialized dictionaries obtained and general dictionaries, which is applied to text processing in social security audit, acquiring corpus of the field. The experimental results show that the proposed method can well divide, extract and discover the conceptual knowledge of field.