Research On The Key Technologies Of Personalized Semantic Query For Movie Information

Watching movies has become a way for more and more people to entertain.Movie information is distributed on different mainstream video websites and these websites cannot provide users with personalized query results,making it difficult for users to search for video resources of their own interest.Aiming at this problem,thesis organizes scattered movie information on the network to build the film information ontology knowledge base.Combining the k-Nearest Neighbor algorithm and ontology query method,analyze users' historical records of watching movies to get the user's interest tendency.First,aiming at the problem of movie information distributed in different sites,thesis analyzes the ways that mainstream websites organize film information and compares different methods of ontology construction.Thesis adopts seven-step method construct the ontology model suitable for the film domain,and designs a method to map the film information data from the source database to the ontology knowledge base.Which provides the film domain ontology knowledge base for the follow-up research.Secondly,thesis investigates four methods of knowledge acquisition to acquire relevant knowledge in the field of film: ontology knowledge reuse,authoritative website interface acquisition,transit website knowledge acquisition and crawler acquisition knowledge.After the comparison of the four methods,thesis selects the crawler to obtain knowledge,and enriches the film ontology knowledge base on the built film ontology model.Then,thesis analyzes the heterogeneous types of film information ontology,formulates the rules of ontology knowledge inspection,and checks the film ontology.Aiming at the problem of incomplete relationship between ontology attributes,thesis adopts the method of ontology reasoning to supplement the ontology knowledge on film ontology knowledge base.Finally,in order to realize the film information retrieval based on the user interest,construct the user interest analysis model,thesis adopts k-Nearest Neighbor algorithm to calculate the weight of the results retrieved from the ontology query based on user interest and compares the results obtained by ontology query and the results retrieved by search engine.