Method for determining sentence similarity

Abstract

The invention discloses a sentence similarity judgment method, which belongs to the technical field of natural language processing. The method comprises: according to two externally input sentence samples, respectively obtaining a character/word vector matrix in each sentence sample; respectively extracting overlapped features in each sentence sample to form an overlapped feature matrix, and combining the corresponding character/word vector matrix with the overlapped feature matrix for each sentence sample to serve as input data of the first neural network model; respectively processing according to the first neural network model to obtain the sentence vector for each sentence sample, then processing the sentence vectors to form a sentence combination vector, and combining the sentence combination vector with an overlapped feature vector formed according to the overlapped features to serve as the input data of the second neural network model; and processing according to the second neural network model to obtain a similarity metric associated with the two sentence samples, and outputting the similarity metric to serve as a basis for determining the similarity of the two sentence samples. The above technical solution has the beneficial effect of providing a method for determining sentence similarity, which can be used for solving the problem that the calculation of sentence similarity heavily depends on the quality of pre-trained character/word vectors and unregistered words in the prior art, thereby improving the measurement method for the calculation of sentence similarity.