Research On Intelligence Information Push Technology For Scientific Personnel

With the rapid development of technology,the output of intelligence information and research results in various fields have increased rapidly,and how to quickly find information that meets the interests of scientific personnel has become an urgent problem.The phenomenon of information overload has accelerated the research of recommendation systems.The intelligence push technology aims to connect scientific personnel with scientific literature,technical reports,news and other intelligence information databases,so that scientific personnel can timely and accurately push intelligence information that meets their needs.Scientific literature and technical reports are one of the important ways for scientific personnel to understand the current technological frontiers and academic trends.In response to the cold start problem of the recommendation system that only has a single perspective of the user scoring information,this paper uses a user portrait with a clear professional background and specific research needs,proposing a personalized recommendation algorithm for intelligence information that integrates user characteristics and semantic computing.On the one hand,using the rich background information and research experience of scientific personnel to generate recommendation result for static information modeling of scientific personnel;On the other hand,user behavior is used to mine the potential semantic relationship between intelligence information for intelligence recommendation.Finally,the two methods are combined.Experimental results show that the recommendation effect of this method is significantly improved compared with the traditional method.The application of auxiliary information can make up for the situation when the user's item interactive behavior is very few.The existing auxiliary information is usually the application of item auxiliary information.This paper attempts to use the user's interactive information as user auxiliary information.Based on the application of collaborative variational autoencoder to item auxiliary information,the auxiliary information with user feature restrictions and intelligence data,which can collaboratively adjust the feature matrix generation of users and intelligence data.The experimental results show that the recommendation based on the bidirectional constraint model has better recommendation performance than the original collaborative variational autoencoder.Finally,intelligence information recommendation system based on user characteristics and semantic computing is designed and implemented,which can implement the personalized recommendation display of intelligence information according to user characteristics and browsing records.