Research And Implementation Of Movie Recommendation System Based On Knowledge Graph

The recommendation system is designed to provide users with accurate item recommendations,which can effectively solve the problem of information explosion caused by the large amount of data,and it is widely used in websites such as movies,shopping,and news.Traditional recommendation systems are limited by data sparsity and cold-start problems,and often cannot make reasonable recommendations.Knowledge graphs can effectively express the semantic relationship between entities,and the application of knowledge graphs to recommendation systems has become a research hotspot today.This dissertation constructs a knowledge graph in the movie domain and obtains the semantic similarity of movies,and obtains the similarity of movie ratings through a collaborative filtering algorithm,and then fuses the similarities between the two to produce a better recommendation effect.The main research contents of this dissertation are as follows:(1)Constructed a knowledge map in the field of film.First,the entities and relationships of the movie knowledge graph are described,and then based on the Movielens data set,the attributes needed to construct the knowledge graph are expanded,and the type of knowledge entities and relationships are determined.Finally,a graph database is established to combine the triple data of entities and relationships.Import it into the graph database and construct a knowledge graph.(2)This dissertation proposes an improved Trans HNK model based on the knowledge representation learning framework.The entity set is divided into multiple clusters through a clustering algorithm,and negative triples are sampled in different clusters.The semantic similarity of the movie is obtained through the knowledge representation learning framework,and then the movie score similarity is obtained through the movie score matrix,and the two similarities are merged to obtain the similarity expression.Through experiments,the hit rate and average ranking of the model are obtained,and the fusion factor of a good recommendation effect is calculated,which ensures that the recommendation algorithm achieves comprehensive and good performance in accuracy and recall.(3)Design and implement the movie recommendation system.This dissertation designs and implements a movie recommendation system based on the knowledge graph-based movie recommendation algorithm.First,it analyzes the requirements of the system,and then designs the system architecture,system functions,database,etc.Then,the movie recommendation interface,movie details interface,The movie management interface and other pages were implemented,and finally passed the test to ensure the stable operation of the system.