Towards automatic multilevel indexing for Thai text information retrieval

To enhance the effectiveness of Thai text retrieval system, we need a significant improvement in automatic indexing. This paper is intended to provide the application of statistical and natural language processing techniques to obtain multilevel content identifiers: phrasal level, single term level and conceptual level. These multilevel indices will cover a very wide range of document retrieval without degradation of system performance. Automatic multilevel indexing for Thai text requires three processes: lexical token identification, phrase identification and extraction, and multilevel index generation. The results give a significant benefit both in precision and recall