Rough集挖掘时间序列的研究

Rouhg集方法是一种用于处理不确定性和模糊性知识的数学工具。探讨了基于Rough集方法的时间序列挖掘问题，提出一种将时态信息系统转化信息系统的方法和一个将实时时态信息系统转换为时态信息系统的方法；并从理论上证明了该方法在挖掘效率上的优越性。

Research of Mining Time Series with Rough Sets

The Rough sets approach is an important mathematical tool to deal with uncertain or vague knowledge. The advantage of dealing with uncertain problem using Rough sets is that it does not need the apriori or extra information of data, and that it is easy to be understood and used. However, because the Rough sets approach has been developed with ordinary non temporal database tables, it is necessary to develop method transforming the temporal information system to traditional information system. In this paper, mining time series with Rough sets is discussed. A method for transforming the temporal information system to the traditional information system is proposed. It lets the attribute value sets of the information system be a set of trends instead of values measured at a certain point in time. Using this approach, dependencies can be traced back in time for as long as we wish. Comparing with the traditional transformation approach, this method can generate fewer attributes in the information system. In addition, a method for transforming the real time temporal information system to the temporal information system is also proposed. It is based on the observation that the less objects in the information system, the more easier Rough set approach being applied to mine time series . This approach lets uniform frequency be minimum value of all (x) for x( U, and can preferably resolve large number of objects in new TIS generated by using traditional transformation approach. An example in this paper demonstrates methods proposed above. Probation shows that these methods can effectively lower computational complexity of mining time series using Rough sets.