基于用户行为的自动任务识别技术研究

Research on automatic task identification based on user behavior

自动任务识别是多任务工作环境下自动任务管理技术的关键,其中对窗口切换历史信息采用Bron-Kerbosky算法来聚类同一任务的窗口,已经被国外研究者采用.然而,该方法仅适用于短时间、较少任务的识别,而对长时间下多个工作任务识别缺乏有效性.本文创新性地提出将窗口切换历史聚类结果与基于焦点时间的窗口重要性相结合形成任务向量,再运用模糊K-Center聚类算法求解任务窗口集合来实现长时间工作环境下多任务识别的方法.实验结果表明,该方法能有效识别长时间工作环境下的多个任务且具有较高的准确率.

Automatic task identification is the key function of automated task management system in multi-tasking working environment. Foreign researchers have taken forward methods based on window switch clustering analysis to calculate task window-application sets. But this method is lack of effective recognition of multiple tasks in long working hours. In this paper, for these limitations, we provide a new method that combines results of window switching clustering with focus-time based window impor- tant rate into mission vector, then use uncertain K-Center clustering algorithm for calculating cluster center, to segment automatic multi-task-identification in long working hours. And, some experiment results are presented and they prove this method improved task classification accuracy obviously.