Mining ordered patterns in sequences

Rules and patterns are two common forms of knowledge representation in data mining. They usually reflect various relationships among objects, among attributes of objects, or between an object and its attributes.; Sequential patterns are important forms of knowledge in the areas of sequence mining, time series analysis, and string algorithms. They usually describe ordered relationships among the objects in a sequence or a collection of sequences. Traditionally, sequential patterns are derived from the subsequences of a sequence or a collection of sequences. Thus, sequential patterns are always composed of the objects from sequences.; This study investigates orders on the domain of sequences. It suggests, based on a weak order, two new forms of sequential patterns. The first form is to add order relations to the traditional sequential patterns. The sequential patterns in this form are represented by the objects and the order relations. The second form of sequential patterns is represented merely by the order relations from sequences. The two forms of sequential patterns are generally called Ordered Patterns. (Abstract shortened by UMI.)...