• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议> >A Method of Group Behavior Analysis for Enhanced Affinity Propagation
【24h】

A Method of Group Behavior Analysis for Enhanced Affinity Propagation

机译：一种用于增强亲和力传播的群体行为分析方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

With the popularity of mobile phones, it is necessary to mine and analyze the user habits, network applications and other data, which can help provide users with a strong adaptability of information services. On the basis of acceleration sensor and touch screen data, we analyze the behaviors of browsing the web, chatting, making calls and playing game. The traditional Affinity Propagation algorithm analyzes all the characteristics of the data as an equal role in group behavior analysis, which has some limitations. In this paper, an Adaptive Feature Weighting based on Affinity Propagation (AFWAP) Group Behavior Analysis Algorithm is proposed, which introduces feature weight into the AP algorithm. The proposed method makes different contribution to the class center in each iteration process, and assigns a new weight for each dimension attribute then to update the feature weight adaptively. In the clustering process, the importance of different features can be measured, which solves the shortcomings of the traditional AP algorithm using equal weight. Finally we apply the proposed method to group behavior analysis.
机译：随着移动电话的普及，有必要挖掘和分析用户习惯，网络应用和其他数据，这有助于为用户提供强大的信息服务适应性。基于加速度传感器和触摸屏数据，我们分析了浏览网页，聊天，拨打电话和玩游戏的行为。传统的“亲和传播”算法将数据的所有特征分析为群体行为分析中的同等角色，这有一定的局限性。提出了一种基于亲和传播（AFWAP）群行为分析算法的自适应特征加权，将特征权重引入到AP算法中。所提出的方法在每个迭代过程中对类中心做出不同的贡献，并为每个维度属性分配新的权重，然后自适应地更新特征权重。在聚类过程中，可以衡量不同特征的重要性，从而解决了传统的等权重AP算法的缺点。最后，我们将提出的方法应用于群体行为分析。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号