| 摘 要: 高校中的社团较多，如何能让大学生快速地找到感兴趣的社团，成为社团管理者面临的主要问题。为了向大学生推荐可能喜欢的社团，本项目通过基于协同过滤的推荐算法，找到共同出现的频率来计算语义的相似度，并通过计算空间向量的夹角余弦值进而计算文本之间的相似度，能够在提高社团管理者工作效率的同时，为需要加入社团的学生和对社团文化有浓厚兴趣的学生提供更加全面的信息。实验结果表明，当推荐项目数量为10时，该方法的召回率、准确率和Fl值分别提高了12.81％、7.65％和14.51％，表明基于协同过滤的推荐算法可有效提高推荐结果。 |
| 关键词: 社团管理；推荐算法；语义相似度；余弦相似度 |
|
中图分类号: TP311.5
文献标识码: A
|
|
| Design and Implementation of College Clubs Recommendation System based on Collaborative Filtering Algorithm |
|
LUO Wei, YIN Hongtao, TAO Chen
|
(Department of Software Engineering, Dalian Neusoft University of Information, Dalian 116023, China)
luowei@neusoft.edu.cn; 208105166@qq.com; 709510257@qq.com
|
| Abstract: There are many clubs in colleges and universities. How to make college students quickly find clubs of their interests has become the main problem faced by club managers. In order to recommend clubs that college students might like, this paper proposes to use a collaborative filtering recommendation algorithm to find the frequency of common occurrences for calculating semantic similarity. The similarity between texts is calculated by calculating the cosine of the space vector angle. It can not only improve the work efficiency of club managers, but also provide more comprehensive information for students who need to join the club and students who are very interested in club culture. Experimental results show that when the number of recommended items is 10, the recall rate, the accuracy rate and Fl value of this method are increased by 12.81%, 7.65% and 14.51% respectively, indicating that the recommendation algorithm based on collaborative filtering effectively improves recommendation results. |
| Keywords: club management; recommendation algorithm; semantic similarity; cosine similarity |
