GENERATING DIGESTS FROM EDUCATIONAL ARTICLES AUTOMATICALLY BASED ON SECOND ORDER HMM

摘要

Automatically generating summary of articles is very important when we encounter explosive reading information; computers can help people on text compression, extraction, representation and obtain core text content automatically. However, computer still encounters a lot of difficulties, for example, how to divide words from ambiguity, inaccuracies, redundancy of the lengthy article, and so on. This paper presents an improved Hidden Markov Model (HMM) Word segmentation method.