摘要
兴趣点推荐已成为帮助人们发现感兴趣点的重要手段,传统推荐算法仅仅对固定时间段内的兴趣点做推荐,而没有考虑兴趣点推荐存在时间上的衰减现象.为了进一步提高兴趣点推荐算法的性能,总结了传统基于兴趣点相似度的协同过滤推荐算法,设计了表现更优异的改进算法——引入时间衰减项的兴趣点推荐算法.实际数据集上的测试结果表明,改进算法在精准度、召回率和F指标等方面比传统基于兴趣点相似度的协同过滤推荐算法有更好的表现.
Point-of-interest(PoI) recornmenda6on has become an important means to help people find interesting locations. Traditional recommendation algorithms just focus the recommendation on a fixed period of time, without taking account of the attenuation phenomena of time. In order to further improve the performance of the recommendation algorithm, this paper summarizes the traditional collaborative filtering recommendation algorithms based on the similarity of PoI and designs a superior improved algorithm which is the PoI recommendation algorithm introducing time attenuation item. The test results on real datasets show that the improved algorithm performs better than the traditional collaborative filtering recommendation algorithms based on the similarity of Pol in terms of precision, recall rate and F-measure.
出处
《杭州电子科技大学学报(自然科学版)》
2016年第3期42-46,共5页
Journal of Hangzhou Dianzi University:Natural Sciences
基金
国家自然科学基金资助项目(61502131)
关键词
兴趣点推荐
协同过滤
时间衰减项
推荐算法
point-of-interest recommendation~ collaborative filtering
time attenuation item
recommendation al- gorithm