Natural language processing method and device combining multiple systems

Abstract

The present invention proposes a multi-system combined natural language processing method. The method includes: extracting the characteristic words in the received text information; The first weight value in each group classified by the keywords; obtain the retrieval results corresponding to the text information from a plurality of third-party systems, and according to each retrieval result, each group has its own subordinate The second weight value in the tripartite system and the first weight value corresponding to each group respectively calculate the scores of each retrieval result, and the retrieval result with the highest score is taken as the output result corresponding to the text information. The present invention can use multiple single natural language processing systems in combination, and multiple systems vote to generate the final answer, so that the results returned by the system are more accurate, and solve the problems of single results, insufficient knowledge, and problems existing in existing natural language processing systems. Questions with too little match between question and answer.

Description

technical field [0001] The invention relates to the field of insurance and finance, in particular to a natural language processing method and device combining multiple systems. Background technique [0002] Existing man-machine dialogue solutions usually return results according to the steps of word segmentation, replacement, and matching. For example, natural language processing systems such as HKUST Xunfei and Turing Robot are usually implemented by maintaining the knowledge base of entries in advance. First replace the question with synonyms, stop words, etc., then segment and match the text, and finally return the entry with the highest matching degree in the database. For some systems, new words and sentence patterns can be learned from the process of question and answer with users, and updated in the knowledge base. [0003] In the human-machine dialogue realized in this way, the robot can only match the results from the existing knowledge base, and the quality of the...

AI Technical Summary

Problems solved by technology

[0003] In the man-machine dialogue realized in this way, the robot can only match the results from the existing knowledge base, and the quality of the answer depends largely on the number of knowledge bases, and it is very easy for the answer to be inconsistent with the question.

[0004] Therefore, in the existing technical solutions, the processing results are single, the knowledge is insufficient, the matching degree between questions and answers is too low, and the processing results are not accurate enough

Images