Semantic recognition and analysis method based on morphological characteristics

Abstract

The invention relates to a semantic recognition analysis method based on language shape features, which comprises the following steps: determining a feature set of questions, and screening through a binary tree and a regular expression to obtain an undetermined language most similar to the semantics of the questions; determining possible answers from a to-be-determined language network, and takingthe subsequent sentences of the to-be-determined language most similar to the questions obtained in the first step as a plurality of possible answers of the questions in the to-be-determined languagenetwork; taking the questions as the input of the semantic analysis neural network, taking the words contained in the plurality of answers as the output, carrying out the forward propagation of the semantic analysis neural network, obtaining the values of the output word nodes, and determining the variables and the final answers according to the output values of the word nodes. According to the semantic recognition and analysis method, language data resources on which algorithm operation depends can be richer, a statement network is denser, and the obtained answers better conform to semanticscenes limited by questions and semantic backgrounds of the questions.

Description

technical field [0001] The invention relates to the field of semantic recognition, in particular to a semantic recognition analysis method based on morphological features. Background technique [0002] With the rapid development of the Internet, semantic understanding technology has been introduced into the dialogue system on smart devices, so that the natural language dialogue system has become a communication tool commonly used by people. [0003] At present, the basic technologies for realizing natural language dialogue systems can be divided into two categories, rule-based methods and statistical-based methods. Among them, the rule-based method refers to using computer language to describe according to the defined grammatical rules, parts of speech and word and sentence construction rules, etc.; the statistical-based method refers to using deep learning and big data to build a dialogue system and automatically generate dialogue. In actual use, the existing dialogue rec...

AI Technical Summary

Problems solved by technology

In actual use, the existing dialogue recognition system is not very capable of recognizing user intentions, and there are often situations where the user cannot answer the user because the user's intention cannot be judged, or the answer is wrong or the answer is repeated, making the dialogue built by the dialogue system difficult. The content is too limited and the user experience is not high

Images