Chinese sentence similarity research based on semantic string

The computation of Chinese sentence similarity has been widely used in many areas. However, due to the complexity of Chinese, currently, there have no corresponding algorithms to solve computation of sentence similarity at home and abroad. This paper represents an improved method of the computation of words similarity firstly, and then on this basis, the structure of semantic string has been given and calculating the impact factor of semantic string, lastly, the improved algorithm of computation of sentence similarity has been presented based on it. The experiment of a certain data sets shows that the improved algorithm is effective.