An improved semantic entropy maximum model and its applications

This paper brings forward a novel semantic discovery approach based on the SEM (semantic entropy maximum). This approach applies the entropy maximum algorithm to the matching and the recognizing of the semantic service, and meanwhile, it compares this algorithm with other related algorithms. The results of those experiments show that the SEM (semantic entropy maximum) model has some preferable performances. In a practical enterprise application of Chinese semantic recognition, the SEM model is proved to be useful and valuable in semantic recognition of Chinese.