Research on Tibetan-Chinese Cross-language Semantic Retrieval Based on Word2Vec

With the increasing of multilingual information on the Internet, it is difficult for people to effectively interpret relevant resources when acquiring information. How to obtain the multilingual information resources needed by users from large-scale information has become an urgent matter, so cross-language semantic retrieval has become an important research direction of information processing technology. In order to solve the performance problems in cross-language semantic retrieval, this paper proposes a method of Tibetan-Chinese cross-language semantic retrieval based on Word2Vec. First, use the word vector to construct the Chinese semantic feature vector of the event keyword, then calculate the feature translation vector of the Tibetan event keyword, and finally complete the cross-language keyword alignment by calculating the similarity between the semantic feature vectors, thereby realizing the key to query automatic translation of words to complete cross-language semantic retrieval. Experiments were conducted on the constructed corpus and the effectiveness of the method was proved.