Research On Knowledge Recommendation Of Digital Library Based On Scientific Research Context

Knowledge recommendation can solve the problem of recommending correct knowledge to knowledge demanders,and get good feedback in a large number of user-oriented recommendation service systems.At present,when digital libraries provide knowledge recommendation for scientific research users,the traditional knowledge recommendation system only considers the basic information,logs and historical records of users,which results in information overload.Situational awareness technology supports the information recommendation system,and can obtain the situation information more comprehensively and actively.It can solve the incomplete information obtained by the traditional recommendation system,which leads to the deviation between reasoning user needs and actual behavior preferences.Therefore,this paper introduces the mature context-aware technology,analyzes the scientific research environment and constructs a scientific research situation model,and then proposes a hierarchical filtering strategy for digital library knowledge active recommendation based on the scientific research situation,which can reduce the information overload phenomenon in the process of scientific research personnel knowledge recommendation.To achieve accurate and active recommendation of digital library knowledge to researchers.Firstly,this paper reviews the development and research status of situational awareness in literature analysis,gives the definition and connotation of scientific research situation based on the actual research background of this paper,and obtains from four dimensions of scientific research personnel,scientific research process,scientific research task and scientific research object according to "scientific research object completes scientific research task in scientific research process".The elements and attributes of scientific research situation are analyzed,and the interaction between the elements of scientific research situation is analyzed.Based on this,the UML object interaction model of scientific research situation is constructed by object-oriented modeling method.Secondly,the demand of digital library knowledge oriented to the whole life cycle of scientific research process is analyzed with the scientific research situation.IDEF0 method is used to analyze the requirement of digital library knowledge in different stages of scientific research process and construct the data structure of scientific research activities.To achieve accurate and efficient knowledge sharing and reuse.Finally,based on the research situation model,this paper proposes a hierarchical filtering recommendation method for digital library knowledge.The strategy combines context prefiltering and situational modeling.According to the situation information,the category of scientific research situation that scientific research users are currently engaged in is deduced.According to the degree of similarity between calculation and digital library knowledge application situation,the required knowledge items for scientific research personnel are preliminarily selected.According to the integrated knowledge screening algorithm of collaborative filtering model,current situation model and historical situation model,the second selection is made.Select the digital library knowledge that researchers need.Based on the above research methods,this paper applies the knowledge recommendation method proposed in this paper to the research situation of a graphic information research team to verify the feasibility and effectiveness of the model and method.With the help of mature situational perception theory,this paper constructs the current scientific research situation model of scientific researchers,so as to obtain more comprehensive information of scientific research situation.Based on the above research contents,this paper proposes a hierarchical filtering knowledge recommendation method for digital libraries,which can effectively achieve the goal of accurately recommending knowledge for researchers.The work in this paper provides a new direction for object-oriented recommendation strategy.