从事件序列中发现一般阶段过程模型

Finding the General Stage Process Models from Event Sequences

生活中存在很多随时间推进而形成的事件序列,比如用户在购物网站上随时间而进行的浏览事件,以及病人随时间而呈现出的各种病症.如果能从这些海量的事件序列中发现规律,找出具有代表性的事件序列模型,并能合理推断序列中的各个独立阶段过程事件,这将具有非常大的理论和实际意义.针对此问题,提出一种从事件序列中发现一般阶段过程模型的方法.首先,详细介绍了本文所用的模型,包括阶段过程模型和事件序列类型模型推断.最后,利用本文所介绍的数据集,通过两大组实验验证了本文模型在预测精度和时间成本方面都具有很好的性能.

Now,there are a lot of event sequences advancing with time, such as user＇s browsing history on the shopping website,as well as the various disorders of patient over time. It will have great theoretical and practical significance that if we can find the law from these massive amounts of event sequences, getting the typical models of event sequence, and can infer the reasonable separate phase processes in a sequence. For this problem,this paper presents a method, which can find the general stage process models from the event sequences. First, the paper introduces the models, including the way of stage process model and event sequence type infer- ence. Finally, using the data sets presented in this paper, the prediction accuracy and the time cost of this mode are well verified by two groups of experiments.