Research And Application Of Recommendation Algorithm Based On Adaptive Clustering And User Trust Model

With the rapid development of the Internet,the data has been increased exponentially.making it more difficult and need spend a lot of time for people to extract useful information from massive data.Massive data acquisition costs are high and overwhelm users,the Recommended algorithm is considered to be the key technology to reduce the user's burden.However,in the face of data sparse,there are still many shortcomings in the efficiency and accuracy of the traditional recommendation algorithm.Therefore,this paper proposes a UTMRA recommendation algorithm,and applies it to student's personalized career planning recommendation system.The main work and innovation are as follows:1.Adaptive Improvement of K-means Clustering Algorithm.AKC algorithm was proposed,aiming at shortcomings of k-means algorithm which the number of clusters to be artificially set and initial center random selection,The algorithm realizes the adaptive selection by finding the first center point through the high density region and using the region restriction strategy.The improved initial center selection and the formation of the number of clusters can be dynamically and adaptively generated based on different data sets,which can significantly improve the aggregation effect.2.UTMRA recommendation algorithm.In this paper,A similarity calculation method and a recommendation strategy based user trust model are constructed,aiming at the limitation of the traditional recommendation algorithm only considering the user history score.In order to improve the accuracy of the recommended results,comprehensively considers the impact of attribute information and common scoring information on recommendation results.3.AKC-UTMRA algorithm.Although the UTMRA algorithm has been improved in accuracy,it has increased the computational complexity of algorithm and reduced the execution efficiency.In order to improve the execution efficiency of the algorithm,AKC algorithm has been introduced.The advantages of AKC algorithm and UTMRA algorithm are fully combined to improve the accuracy and reduce the search space of the UTMRA algorithm.Through comparison experiments,it is shown that both accuracy and real-time performance are improved.4.Occupation planning recommendation system based on AKC-UTMRA algorithm.Supported by related topics,a set of personalized career planning recommendation system was designed and developed based on this algorithm which achieves the expected results in the system,The intelligent recommendation of the student career planning system was completed and the student personality characteristics were automatically matched.