Font Size: a A A

Voice Question Answering System Based On Clothing Recommendation Knowledge Graph

Posted on:2022-07-28Degree:MasterType:Thesis
Country:ChinaCandidate:X LiuFull Text:PDF
GTID:2518306494481084Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Accompanied by the rapid popularity of apparel e-commerce,an increasing number of people tend to shop for clothing online.But people generally lack the professional knowledge of clothing collocation,and only by their own aesthetic or influencer marketing to buy clothing,it has a certain blindness.At the same time,the exponential expansion of data on the web complicates the user's ability to quickly find valuable information in the overwhelming amount of data.Providing personalized clothing recommendation services,improving retrieval efficiency and optimizing user experience have become the demands of the majority of users.In view of the above problems of clothing e-commerce,this topic through the research image design expert rule system,constructed the knowledge graph of clothing recommendation,and proposed a clothing recommendation scheme based on the rules of clothing recommendation.Then,personalized clothing recommendation is realized.At the same time,a voice question answering system based on the knowledge graph of clothing recommendation has been designed and developed.It combined the voice technology and intelligent question answering technology to optimize the user experience,which improved the retrieval efficiency of users.The main contents of this paper are as follows:Firstly,a comprehensive sorting and formatting of image design specialist knowledge and theories related to the apparel field has been carried out,an image design expert rule system was constructed,expert rules were processed using knowledge graph technology,and the apparel recommendation knowledge graph was stored in a Neo4 j graph database.Besides,a recommendation method based on the apparel recommendation knowledge graph is proposed.Then,the paper designed the voice question and answer process based on the knowledge graph of clothing recommendation,which was divided into two modules: question processing and sentence generation.The question processing module trained BERT-Bi LSTM-CRF entity extraction model for named entity recognition,the results were processed by template matching.Finally,based on the above research,a voice question answering system based on the knowledge graph of clothing recommendation was developed.The system was developed by applying the Python language,based on the Flask framework,calling on the i FLYTEK voice cloud service to realize voice interaction,and the visualization of clothing recommendation knowledge map was realized by using Echarts,an open source chart library of Baidu.This paper introduces in detail the methods to deal with the expert knowledge of image design,the construction process of clothing recommendation knowledge graph and the development process of voice question answering system.Through the system verification,the results have shown that the voice question answering system has good performance in retrieval efficiency and clothing personalized recommendation,and has a certain application value.
Keywords/Search Tags:clothing recommendation rules, knowledge graph, named entity recognition, voice question and answer
PDF Full Text Request
Related items