Knowledge graph-based interactive question and answer method and system

Abstract

The invention discloses a knowledge graph-based interactive question and answer method and system. The method comprises the steps of constructing a knowledge graph, wherein data in the knowledge graphis from multiple open-source information sources; according to existing entities in the knowledge graph, forming a dictionary, forming mapping from a name to a professional field through a manual tagging method, and performing expansion in a conventional feature modeling mode to form a professional dictionary; according to the data in the knowledge graph, forming mapping from the entities to a training set of the field through the manual tagging method to establish a classifier; according to the professional dictionary, performing word segmentation on a natural question sentence through a forward maximum matching method, and according to results after word segmentation, inputting the results to the classifier for performing classification, thereby classifying the natural question sentenceinto questions in different fields; and mapping the classified questions to obtain corresponding question templates, and converting the question templates into query sub-graphs in the knowledge graph. Answer can be performed for sentences of more complex questions, so that the answer quality can be ensured and the manual intervention is effectively reduced.

Description

Technical field [0001] The present invention relates to the field of artificial intelligence technology, and more specifically, to an interactive question and answer method and system based on knowledge graphs. Background technique [0002] Traditional automatic question answering systems are divided into two parts: question processing and answer retrieval. The basis of question sentence processing is word segmentation. However, using word segmentation to process long text nouns in some professional fields can easily cause the separation of nouns. The common idea to deal with this problem is to manually construct a professional dictionary. This method consumes a lot of human resources. [0003] Secondly, in terms of answer retrieval, traditional question answering systems use a scoring mechanism to select a series of candidate answers from massive text data; and then build a selection function again to select the closest answer from the candidate answers. The disadvantage...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the defects of the prior art, the purpose of the present invention is to solve the problem that the existing automatic question answering system is easy to cause division of nouns when dealing with long text nouns in some professional fields, and adopts the construction scoring mechanism to select a series of candidate answers from massive text data , the selected answer will have different degrees of wrong technical problems due to the different mechanisms and functions of the construction

Images