Research On Optimization And Application Of Fusion Text Information Association Matching Mode

The rapid development of computer technology has brought convenience to users to obtain massive information easily,but also increased user’s burden on processing massive information.Therefore,how to classify and organize text information resources so that users can retrieve the target resources more conveniently is a problem in current information management and information services.To achieve the following goals: quickly retrieve the appropriate text information resources for the users from the text information collection;complete the User-resource association matching model;make the User-resource association matching more accurate;and solve the problem that text infromation is not fully utilized in infromation resource management;I propose a User-resouce association matching model integrating text information.The User-resource association matching model integrating text information is studied from two directions: resource semantic feature extraction and user-resource interaction feature extraction.Then based on real data on the loans of books,application of library book recommendation service is realized.The specific research work and results are as follows:(1)Analysis and optimization of text information resource matching.I have studied Chinese text vector in depth by natural language processing.For the current problem of insufficient meaning of text vector,the semantic information of two word vectors is obtained through the fusion of word vectors generated by BERT and Glo Ve model,so that the semantic information representation ability of the fusion word vectors of the fusion word vectors in the corpus of small data is enhanced.The shallow text convolutional neural network is used to extract the features of the fused word vector,so that the fused word vector can accurately express the meaning of words,and also reduce the hardware requirements and deployment difficulty in the experiment.Compared with the traditional Glo Ve model and a single BERT pre-training model,the resource matching accuracy of the fusion word vector model also can be improved with less sample data.(2)The study of User-resource association matching model integrating text information,which relies on historical behavior information to match User-resource,wasting effective text information in resources.So this paper proposes a User-resource association matching model integrating text information,which combines User-resource behavior information with the text information of the resource itself,and produces the final association matching coefficient according to the association matching index of the resource to the target user.Through experimental comparison,the User-resource association matching model integrating text information not only improves the reliability and stability of the book recommendation system,but also significantly improves the accuracy of book recommendation.(3)Application research of association matching model.Based on the User-resource association matching model integrating text information,this function can be applied to the library book recommendation service.Then the Tkinter pop-up window can be used to realize user interaction.Finally achieve the task of book recommendation to the target reader.