摘要
对推荐算法进行综述和分析,针对目前推荐方法的用户兴趣不明显,针对性较差等问题,提出一种基于访问时间、资源种类和心情留言的推荐算法。其中,心情留言用于衡量用户喜爱资源的程度,将该算法命名为TTM(Time-Types-Mood message)算法,并提出基于访问时间、资源种类和心情留言的三种数据权重函数。该算法在学习系统中用于对用户行为进行分析。实验证明,这种TTM算法能够做出合理的推荐,推荐质量得到了提高。
For the user interest is not obvious, targeted poor and other issues in the current recommendation methods, this article reviews and analyzes the recommendation algorithms, and a recommendation algorithm based on the access time, resource type and mood message is proposed. Among them, the mood message is used to measure the extent of the user＇s favorite resources. The algorithm is named TTM （Time-Types-Mood message） algorithm, and three data weighting functions is proposed based on the access time, resource type and mood message. The algorithm is applied to the learning system to analyze user behavior, and make verification to itself. The experimental results show that TTM algorithm can make a reasonable recommendation; the recommendation quality has been improved.
出处
《计算机时代》
2015年第11期4-7,共4页
Computer Era
关键词
用户行为分析
推荐算法
学习系统
数据权重
数据挖掘
user behavior analysis
recommendation algorithm
learning system
data weight
data mining
