| With the informationization and networking of social life,the increase of information content will inevitably lead to the filling of useless garbage content,resulting in the reduction of effective information ratio.In order to solve the problem that traditional search methods cannot provide effective information accurately under the condition of huge data and unknown demand,recommendation system emerges from this.According to the different preferences of users,it can provide more accurate information services for different users’ personal needs.Recommendation systems infer users’ thoughts and push the content they may be interested in,so that the way of Internet information service changes from passive to active.At present,various professional and technical fields are in the trend of continuous integration,which brings great convenience for some users to find some books that are not familiar with the professional field.In the face of the problems existing in searching books in university libraries,it is more urgent to apply the recommendation algorithm to the management of professional books.Therefore,the problem to be solved in this paper is to integrate the recommendation algorithm into the traditional library management system.The main task of this paper is to build the system based on The Hadoop platform and select Mahout’s collaborative filtering recommendation algorithm as the core algorithm to complete the collection recommendation system of university library based on Mahout.The main work of this paper is as follows:First of all,the realization principles of the two collaborative filtering algorithms based on users and projects are deeply analyzed and studied,and the application scenarios,advantages and disadvantages of the two collaborative filtering algorithms are compared,and the advantages of the two algorithms are reasonably utilized in combination with the characteristics of the system to achieve the goal of complementary advantages.Secondly,the author thinks about the three problems existing in the recommendation algorithm and analyzes the causes of the problems.Combined with the characteristics of the system,the author proposes the optimization of the cold start of new users and new projects,the improvement of data sparsity and the improvement of performance.Finally,analysis of the requirements of the system,design system functions.This system uses SSM framework and My SQL development commonly used in Java development to realize data interaction between background,foreground and database.In the implementation of the recommendation algorithm,Mahout framework is used,and then on this basis,the system is extended to complete the design and implementation. |
