×

Finding informative rules in interval sequences. (English) Zbl 1029.68755

Hoffmann, Frank (ed.) et al., Advances in intelligent data analysis. 4th international conference, IDA 2001, Cascais, Portugal, September 13-15, 2001. Proceedings. Berlin: Springer. Lect. Notes Comput. Sci. 2189, 125-134 (2001).
Summary: Observing a binary feature over a period of time yields a sequence of observation intervals. To ease the access to continuous features (like time series), they are often broken down into attributed intervals, such that the attribute describes the series’ behaviour within the segment (e.g. increasing, high-value, highly convex, etc.). In both cases, we obtain a sequence of interval data, in which temporal patterns and rules can be identified. A temporal pattern is defined as a set of labeled intervals together with their interval relationships described in terms of Allen’s interval logic. In this paper, we consider the evaluation of such rules in order to find the most informative rules. We discuss rule semantics and outline deficiencies of the previously used rule evaluation. We apply the J-measure to rules with a modified semantics in order to better cope with different lengths of the temporal patterns. We also consider the problem of specializing temporal rules by additional attributes of the state intervals.
For the entire collection see [Zbl 0977.68809].

MSC:

68U99 Computing methodologies and applications
68T05 Learning and adaptive systems in artificial intelligence

Software:

SPADE
PDFBibTeX XMLCite
Full Text: Link