Frequent Temporal Inter-object Pattern Mining in Time Series

Abstract

Nowadays, time series is present in many various domains such as finance, medicine, geology, meteorology, etc. Mining time series for useful hidden knowledge is very significant in those domains to help users get fascinating insights into important temporal relationships of objects/phenomena along the time. Hence, in this paper, we introduce a notion of frequent temporal inter-object pattern and accordingly propose two frequent temporal pattern mining algorithms on a set of different time series. As compared to frequent sequential patterns, frequent temporal inter-object patterns are more informative with explicit and exact temporal information automatically discovered from many various time series. The two proposed algorithms which are brute-force and tree-based are efficiently defined in a level-wise bottom-up approach dealing with the combinatorial explosion problem. As shown in experiments on real financial time series, our work can be further used to efficiently enhance the temporal rule mining process on time series.