基于邻居选取策略的人群定向算法

发布时间：2018-01-25 13:03

本文关键词： 种子人群行为相似人群邻居选取策略用户相似度人群定向　出处：《计算机研究与发展》2017年07期 　论文类型：期刊论文

【摘要】：人群定向是广告推荐系统中的一种重要技术,它是通过分析种子人群的行为数据,找出潜在的目标人群,而现有人群定向算法大多依赖于传统的协同过滤推荐算法.由于传统的协同过滤算法具有推荐精度低和抗攻击能力较弱的问题,为了解决这些问题,提出了一种基于邻居选取策略的人群定向算法.1)通过用户行为相似,动态选择出与种子人群具有相似行为的用户;2)以用户特征和用户行为作为邻居选取的依据,通过用户相似度从行为相似人群中选择出每个种子用户的邻居,并将所有的相似邻居作为候选人群;3)通过基于邻居选取策略的人群定向算法,从候选人群中择出潜在的目标用户,以完成人群定向.实验结果表明:与现有方法相比,该方法不仅提高了人群定向的精度,而且也增强了系统的抗攻击能力.
[Abstract]:Crowd orientation is an important technology in advertising recommendation system. It can find out the potential target population by analyzing the behavior data of seed population. Most of the existing crowd orientation algorithms rely on the traditional collaborative filtering recommendation algorithm. Because the traditional collaborative filtering algorithm has the problems of low recommendation accuracy and weak anti-attack ability, in order to solve these problems. A crowd orientation algorithm based on neighbor selection strategy. 1) dynamically selecting users with similar behavior to seed populations through user behavior similarity; 2) based on the user characteristics and user behavior as the basis of neighbor selection, the neighbors of each seed user are selected from the behavior similarity group by user similarity, and all similar neighbors are selected as candidate groups; 3) through the crowd orientation algorithm based on neighbor selection strategy, the potential target users are selected from the candidate population to complete the crowd orientation. The experimental results show that: compared with the existing methods. This method not only improves the accuracy of crowd orientation, but also enhances the anti-attack ability of the system.
【作者单位】：武汉大学计算机学院;汉口学院计算机科学与技术学院;
【基金】：国家自然科学基金项目(61272277)~~
【分类号】：TP391.3
【正文快照】： 目前随着互联网的发展,互联网广告的数据资源以惊人的速度增长,为了解决广告主向目标用户投放广告的需求,人群定向技术应运而生.人群定向主要通过现有用户的行为,找出未来潜在的目标人群,并选择适当的广告投放给这些目标人群.人群定向的精度依赖于信息过滤的技术,而在现有的技

【参考文献】