Mining rare sequential patterns in data streams with a sliding window
Weimin Ouyang · 2016
Sequential pattern is one of most important topics in researching on data mining and knowledge discovery. Traditional algorithms for mining sequential patterns have two limitations, the first one is only frequent sequence to be considered, and the second one is restricted to the static environment of data set. However, some infrequent patterns can also uncover very interesting knowledge from the data set such as rare sequential pattern. To my best knowledge, current researches on rare sequential patterns are limited to the static database environment, and there is no research work for mining rare sequential patterns over data streams. The author propose an algorithm for mining rare sequential patterns over data streams with a slide window in this paper. Experiments on the synthetic data stream shows that the proposed algorithm is efficient and scalable.