Construction And Application Of Knowledge Graph Of Women's Clothing In The Field Of E-commerce

With the development of the Internet industry,residents' consumption patterns have gradually shifted from offline consumption to online consumption,and the commodity data and user data of e-commerce platforms have experienced large-scale growth.Women's clothing is a representative of individuality and fashion trends,and has unique product characteristics.The knowledge graph uses the graph structure to organize commodity resources,establishes connections between users and commodities,and provides thoughts and opinions for the development of subsequent commodity recommendation and knowledge reasoning.This article takes women's clothing products in the field of e-commerce as the research object,and conducts research on the construction and application of commodity knowledge graphs.The main research methods and contents are as follows.First,this paper uses the literature analysis method to sort out the research status of knowledge graphs,and determines that the research content of this paper will focus on three important modules: data acquisition and preprocessing,information extraction,and knowledge fusion.Second,construct the model layer of the commodity knowledge graph,define the entities and relationships of commodity information resources,and use the method of information transformation to convert semi-structured commodity detail information into triples to store the knowledge graph.Third,a set of manual labeling specification system is designed for the characteristics of women's clothing products,and the deep learning model of BERT+Bi LSTM+CRF is used to identify the entity of product features and emotional tendencies from unstructured user review data.Fourth,firstly integrate the entity category characteristics and the dependency syntactic structure characteristics,and use the rule template to extract the relationship.Then Word2 Vec is used to construct the entity word vector matrix,and the cosine similarity is used to calculate the entity similarity to complete the entity fusion.Finally,the entities that have undergone entity fusion are constructed into a triple structure of and ,and imported into the product knowledge graph through commodity links to realize the product knowledge graph Application expansion.The results show that this article takes women's clothing as an example.On the one hand,it completes information conversion for semi-structured product detail information and constructs the main structure of the product knowledge graph;on the other hand,it completes entity relationship extraction,entity fusion,and entity fusion for unstructured user review information.Knowledge graph storage,in which the F1 value of entity recognition reaches 85.61%.According to the above two steps,this article completes the construction of the knowledge graph of women's clothing in the field of e-commerce,and explores the application of the commodity knowledge graph from the perspectives of information retrieval and satisfaction evaluation.