Research On Personalized Recommendation System In The Digital Library Based On Knowledge Context

The greatest feature of the digital information age is omnipresent, the rapidgrowth of information lead to information overload unlimited digital library as themain carrier of information, also facing the same problem.The personalized recommendations based on user interests, browsing history andother information for the user to find the most interesting information, suchpersonalized service allows users to information from passive viewers to becomeactive participants, information providers also provide passive became active pushinformation, such information providers and users to create a more favorablerelationship, increase user loyalty and reliability.In this paper, based on collaborative filtering techniques, focusing on situationalelements added knowledge in the collaborative filtering technology to constructknowledge through situational model ontology, the user and the user to calculate thesimilarity model of the project and the project context, the use of this article toimprove the Top-N algorithm, the highest similarity to the user recommended N items.The main work is as follows: First, the collaborative filtering algorithm detailedanalysis done research on algorithms emerging data sparsity problem and propose anew and improved way; On the basis of the ontology, users and project constructionfeature model and add context books and readers come to a knowledge-based contextmodel features books and readers; Paper introduces the concept of "HowNet"sememe similarity to optimize the word similarity calculation method, by optimizingthe rule calculation method to calculate word similarity between users, situationalsimilarity between the project and the project, using the algorithm recommendedTop-N N most similar projects recommended to the user. Finally, the experimentsshow that the proposed method improves the efficiency of personalizedrecommendations to some extent.