Ozden, B.; Ramaswamy, S.; and Silberschatz, A. 1998.  Home/Search   Document Not in Database   Summary   Related Articles

....shows that data cube structure provides an efficient and effective structure not only for on line analytical processing (OLAP) but also for on line analytical mining. Notice although the problem of mining segment wise periodicity shares some similarity with that of mining cyclic association rules (Ozden, Ramaswamy, Silberschatz 1998), our study is not confined to perfect cyclicity but allows periodicity with certain confidence. Furthermore, the introduction of data cubes and bit map techniques improves the efficiency of periodicity mining. The remaining of the paper is organized as follows. In Section 2, concepts related to ....

....one may put the algorithm in a for loop and set the for loop index as the period p from 1 up to the entire length of the time series. This technique, though straightforward, may require a lot of processing power. If the confidence threshold is 1 (perfect periodicity) many properties explored in (Ozden, Ramaswamy, Silberschatz 1998) can be adopted to reduce the search effort. For example, if a time series is periodic with a period of p, it is periodic for any multiples of p. This implies that one may search for segment wise periodicity from small periods up, cross out all the multiples of a period along the way to reduce the ....

Ozden, B.; Ramaswamy, S.; and Silberschatz, A. 1998.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC