Abstract
This paper puts forward a multilevel algorithm of Chinese sentence similarity on the basic of predecessors' research. The algorithm considers disordered similarity and ordered similarity: disordered similarity is obtained by computing the ratio of synonyms after Chinese segmentation; ordered similarity is got by calculating the corresponding similarity of component location words. Meanwhile, the sentence pattern conversion is also considered, and if it be a Chinese question, sentence mood parameter need to be used to modify the result. Finally the algorithm rationality is verified by contrast experiment.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have