基于用户身份特征的多标签分类算法

发布时间：2018-02-01 02:35

本文关键词： 社会网络智慧校园启发式规则多标签判断集成学习　出处：《计算机应用》2017年06期 　论文类型：期刊论文

【摘要】：目前对于智慧校园中的家校沟通,缺乏一种衡量和参考的方法。针对智慧校园中特有的聊天特点即存在明显的身份特征,提出了一种基于用户身份特征的多标签分类算法——Adaboost.ML。首先,新增加了启发式规则;然后,引入Adaboost.MH算法,同时摒弃了把数据集进行分片的概念;最后,直接利用单条数据作为分析的焦点,减少了由于时间片边缘带来的误差和推断时间,综合决策出聊天用户之间的关联关系。实验结果表明,与基于规则的启发式方法相比,所提算法在智慧校园数据集上的误报率、漏报率分别降低了53%、66%,同时在微信数据集上也具有良好的分类效果。该算法已应用到智慧校园项目中,能够迅速并准确地了解到家校沟通的情况。
[Abstract]:At present, there is a lack of a measure and reference method to communicate between home and school in the intelligent campus. In view of the unique chat characteristics in the intelligent campus, there are obvious identity characteristics. This paper proposes a multi-label classification algorithm based on user identity, Adaboost.ML.Firstly, a new heuristic rule is added. Then, the Adaboost.MH algorithm is introduced, and the concept of slicing the data set is abandoned. Finally, the single data is directly used as the focus of the analysis, which reduces the error and inference time brought by the edge of the time slice, and synthetically decides the relationship between chat users. The experimental results show that. Compared with the rule-based heuristic method, the false positive rate of the proposed algorithm on the intelligent campus data set, the false alarm rate is reduced by 53% or 66% respectively. At the same time, it has good classification effect on WeChat dataset. The algorithm has been applied to the intelligent campus project, and can quickly and accurately understand the situation of home-school communication.
【作者单位】：湖南大学信息科学与工程学院;湖南大学可信系统与网络实验室;
【分类号】：TP301.6
【正文快照】： 0引言在社会信息化的大背景下,构建“智慧型”校园,利用学校为主体的教育信息化推进过程,成为教育信息化的重要组成部分[1]。智慧校园中家校沟通模块充分体现了学校与家长之间的互动和联系,通过对沟通情况的分析能够有效地反映用户之间存在的本质关联。本文研究把聊天用户之间

【相似文献】