Recommend System Diversity Research And Its Application In Employment Recommendation

Employment recommendation system has a good effect in solving the problem of employment,so it has attracted wide attention by scholars both at home and abroad,which has achieved rich results.However,in the field of employment recommendation,the following deficiency still exists for further improvement: First,the recommendation result is monotonous,so user view is limited.Second,hot jobs are recommended to too many job seekers,which reduces the success of finding a job.Third,unpopular jobs can't get effective recommended,which is harming the interests of employers.Based on the employment recommendation diversity optimization as the main target,we make a thorough research for the problem.The main innovation points and contributions in this paper are as follows:(1)In view of the problem of individual diversity in recommendation system,a kind of individual diversity optimization recommendation algorithm is put forward based on clustering technique.For the reason that recommendation algorithm based on the traditional employment is lack of respect for the individual diversity,resulting in recommendation result is monotonous,a kind of individual diversity optimization recommendation algorithm based on clustering is put forward in this paper.First of all,the algorithm calculates the dissimilarity of items in the system,and takes into account the differences among the item attribute values in the calculation.Secondly,k-means clustering algorithm to cluster items is used according to the item dissimilarity.Then,obtain the predictive score array by the existing recommendation algorithm,set the score threshold,filter items that the predicted score is greater than the threshold,and build user candidate commendation list.Finally,combining with the project clustering information from the user list of candidates,a set of diversity of good projects is obtained and recommended to users.The experimental results show that the algorithm can effectively improve the recommendation of individual diversity while maintaining an acceptable level of recommendation accuracy.The individual diversity of clustering optimization recommendation algorithm applied in the field of employment recommendation can effectively improve the individual diversity of employment recommendation and user satisfaction.(2)In view of the aggregate diversity problem in the recommendation system,an improved algorithm for recommendation aggregate diversity is put forward based on the bipartite graph networks.According to lack of overall diversity considerations on the traditional employment recommendation algorithm,"Matthew effect" caused to the system increasingly serious,and the "long tail" position number increasing phenomenon,this paper proposed an enhanced algorithm for recommendation aggregate diversity based on the bipartite graph networks.Firstly,obtain the predictive score array by the existing recommendation algorithm,set score threshold,filter items that the predicted score is greater than the threshold,and build recommendation list.Secondly,construct the recommendation bipartite graph based on user candidate recommendation list.Finally,based on the recommendation bipartite graph,adopt changing the matching path and the unmatched path to improve the aggregate diversity of the recommendation.The experimental results show that the proposed algorithm can effectively guarantee the accuracy of the recommendation results as well as improve the aggregate diversity of the recommendation.Application of the new algorithm based on the global optimization of the aggregate diversity of the bipartite graph network to employment recommendation area can effectively improve the overall diversity of employment recommendation and user satisfaction.(3)Based on the above two kind of various optimization strategy,a prototype system of employment recommendation is realized.