摘要
随着智能家居行业的蓬勃发展,用户对智能家居所能提供的智能化服务需求越来越大.而现有的智能家居系统通常只能按照预先设定的控制方式和规则简单重复运行,不能根据用户的使用习惯来适时适度的推荐能够反映其个性化需求的控制策略.针对这种情况,采用主流数据挖掘方法来预测用户个性化行为.基于云合智能家居系统开展了相关实验,模拟了一年内10个家庭日常数据.经过对三种数据挖掘算法的对比实验,所采用的支持向量机算法能为不同用户提供符合其个性化需求的服务.
With the flourish of smart home trade, users are looking forward to the services it can provide. The existing smart home sys- tems usually can only run repeatedly according to the preset control mode and rules. But it cannot timely and moderately recommend the control policy which can reflect their individual demand, according to the users＇ habits. Aiming at this situation, this paper used the mainstream data mining methods to predict users＇ individual behavior. Based on Yunho smart home system,we conduct a series of rel- evant experiments. And we simulate ten families＇ daily data in a year. By comparing the experimental results of three kinds of data mining algorithm, it has proved that the data mining methods of this paper can provide different users with services which meet their individual demand.
出处
《小型微型计算机系统》
CSCD
北大核心
2015年第12期2794-2797,共4页
Journal of Chinese Computer Systems
基金
国家科技支撑课题项目(2013BAD19B06)资助
