A time series representation for temporal Web mining

Mining data collected over time without ignoring its temporal information is highly needed for the advance of the information society. The awareness of the need for treating temporal data as an ordered collection of events is increasing. However, using temporal data as temporal sequences without any preprocessing fails to retrieve relevant features of the corresponding data. Therefore, before applying mining techniques, an appropriate representation of temporal sequences is required. Our representation of time series aims at improving the possibility of specifying and finding a significant event. In our new approach, we propose to use data band ranges and areas in order to determine the importance or the weight of a segment. According to the closeness of a segment to a data band range, this new representation can help to discover a significant occurrence. Our representation of time series can be used in different domains, including aviation science and earth science, and can also be applied to, for instance, temporal Web mining. This paper emphasizes on our representation of time series