Research On The Realization Of Ningxia Characteristic Food Intelligent Customer Service System By Text Similarity

Online shopping is almost integrated into every aspect of everyone's life with the characteristics of rich and diverse items,convenience and low price.The order quantity of online shopping platform is increasing rapidly,which requires a large number of customer service personnel from businesses to meet the needs of users.The introduction of intelligent customer service effectively relieves the pressure of customer service personnel.At present,although some platforms have access to the independent Q&A system,there is always a situation of giving an irrelevant answer.This paper focuses on text similarity.By improving the accuracy and F value of similarity algorithm,it makes the intelligent customer service system of Ningxia featured food more intelligent.The specific work is as follows:(1)In order to improve the calculation efficiency,the vector space model only pays attention to the word frequency,ignores the order of the characteristic words,does not consider the specific meaning of words and the grammatical structure of the text,resulting in the loss and distortion of the text content.This paper proposes an algorithm of semantic similarity fusion,which combines the semantic similarity calculation and the spatial vector model,sets a reasonable weight coefficient,and balances the single one of the shortcomings of the algorithm is to optimize the accuracy of similarity calculation.(2)In the calculation of semantic similarity based on HowNet,the similarity only depends on the adjustment parameter α and the distance length in the tree semantic hierarchy.In view of the strong subjectivity of the value of α,and the problem that only using semantic distance as the measurement standard of similarity is not comprehensive,this paper combines the calculation formula of semantic similarity proposed by Liu Qun,and integrates three decisive factors:semantic depth,public semantic distance and semantic dissimilarity,and improves the calculation method of semantic similarity.Finally,by analyzing the experimental data,compared with the single algorithm,the accuracy and F value have made some progress.The results show that this method improves the effect of similarity calculation to a certain extent.