Research On Application Of Collaborative Filtering Recommendation Algorithm In Educational Administration System

With the explosive growth of information in the era of Internet big data,people gradually evolved from the era of information scarcity to the era of information overload.The increase in the amount of information has also greatly changed people's lifestyle.In the early period,the difficulty of query collection caused by information overload has increased.In order to solve this problem,users are able to obtain valuable information more easily,the traditional data retrieval and other traditional methods can not meet the demand,and the relevant recommendation system at this time.Came into being.The recommendation system can find out in a large amount of information that the user may be interested in recommending to different target users.The recommendation system has been applied to various industries and fields in life,and has played an important role in real life.It has brought a good experience to users and has significantly improved the efficiency of applications.After the university's educational administration system has gradually increased its amount of information,it may consider using relevant recommendations to implement some of its functional businesses.For example,the elective module in the university's educational administration system may cause students'selection of course information to be inconvenient during each semester course selection.At this time,the recommendation system can be added to the student's elective course,and different courses can be personally recommended to the students.The efficiency of elective courses,while escalating the interactive experience.The purpose of this paper is to study the application of relevant recommendation algorithms in the elective system of academic affairs.This study will implement personalized recommendation of students to elective courses and recommend them to similar students according to the preferences of certain students,so that the educational system can change traditional users.Interactive ways to improve student elective efficiency.The recommendation system has a wide range of applications in real life.Among them,there are many application models and scenarios of collaborative filtering recommendation algorithms in the recommendation algorithm.After students join the class selection system,the efficiency of student selection will increase and the interaction will change.Under the traditional course selection mode,students can only achieve course selection by searching.The purpose of the recommendation system is to actively recommend more interesting courses to students,not only a new mode of course selection but also a new type of user.System interaction upgrade.Although the collaborative filtering recommendation system will bring about many optimizations and changes,the collaborative filtering recommendation algorithm also has many deficiencies,such as system cold start problems,data sparsity issues,recommendation system portability issues,etc.These are all recommended systems need to overcome difficulties or problems.At the same time,the system calculates user similarity may encounter some popular projects(many students choose courses)affect the final result,in order to make the final recommendation results personalized,in the process of calculating user similarity by adding a penalty factor,aimed at reducing popular projects the heat,so that it can avoid the final recommendation results to a certain extent,improve the accuracy of the recommendation results.This article uses the collaborative filtering recommendation algorithm and k-means clustering algorithm to calculate the recommendation results.Collaborative filtering recommendation algorithm is based on user and model-based two main recommendation algorithm models,and the final recommendation result accuracy rate is directly related to the recommendation model(similarity calculation function)and parameter selection(neighbors number,number of recommendation results).On the basis of different recommendation algorithm models,the influence of different parameters on the recommendation effect is studied,and the recommendation algorithm that is more suitable for this paper is analyzed.Based on the previous theory,this paper first validates the recommendation algorithm in the Python environment,and obtains the recommended results under different algorithm models,and analyzes the changes of the different parameter conditions of these recommendation results,and compares the MAE(average absolute error)and Precision respectively.(Accuracy),Recall,and F-measure parameters.After the algorithm is verified,a better recommendation model based on collaborative filtering recommendation algorithm and k-means algorithm is implemented under the Java platform.Finally,the entire recommendation module is deployed to the student selection system.The main innovations of this paper include the following two points:In order to reduce the popularity of hot items when calculating user similarity,a penalty factor of(?)is added to the penalty mechanism;two different algorithm models are mixed to form a new computing model is a hybrid collaborative filtering recommendation algorithm based on k-means clustering.