Based On Cloud Computation Coordination Filtration Recommendation Algorithm In Wisdom Library Application

At present, the wisdom of Library’s digital resources is exponentially increasingnumber of users, the rapid multiplication. In the face of massive digital resource,computing technology can be more intelligent, more efficient will recommend digitalresources to users by using cloud. With the arrival of the era of wisdom library cloud,cloud computing technology for the intelligent library provides a new service modeland provides effective security system.This paper first introduces the related theory of cloud computing and wisdomlibrary related theory, introduces the service model and the open source cloud conceptand deployment model of cloud computing, cloud computing platform Hadoop.Starting from the concept of the wisdom of wisdom library, library service platformconstruction, the recommended process online display and wisdom library aredescribed, then the research focus on user based and item based collaborative filteringalgorithm, puts forward the traditional user based and item based collaborativefiltering recommendation algorithm problems, solutions and the performance of thetraditional collaborative filtering recommendation algorithm evaluation. Research oncollaborative filtering recommendation algorithm is proposed in this paper by thetraditional combination of multiple users and multiple nearest neighbor algorithmbased on optimal project, the critical point of the plurality of neighbor users andmultiple nearest neighbor projects through the relevant data sets of experimentalmethod, finally built the Hadoop cloud computing platform, using the Hadoop cloudcomputing platform for simulation experiments on user based and item basedcollaborative filtering algorithm. The increase in the number of users in differentconditions, the simulation experiment of similarity calculation method that a pluralityof neighbor users and multiple nearest neighbor project effective combination,comparison of above three kinds of collaborative filtering recommendation algorithmrecommended by performance evaluation, different data sparsity.Therefore, the traditional collaborative filtering recommendation algorithmsbased on user or whether the item based collaborative filtering algorithm, areinevitable problems encountered in data sparse, cold start, development and influencethe performance of computational complexity. Collaborative filteringrecommendation algorithm combined with the nearest neighbor users andneighboring project, the dynamic similarity between users according to specificsituation of nearest neighbor search for users, and the nearest neighbor user itemrating data for searching neighbor project, the experimental results show that theeffective combination of the two, can more effectively adapt to the recommendationsystem, sparse lower better scalability, low time complexity than the traditional collaborative filtering recommendation algorithms, this algorithm can efficiently,wisdom wisdom library books, can satisfy the need of wisdom library.