Estimating temporal pattern bounds using negative support computations

摘要

In many applications and situations, there is a need to estimate which temporal patterns are similar from a temporal database at various timestamps. In such a case, we have to obviously determine true supports of these patterns. If we can somehow reduce number of times, true support is found then it reduces cost of computation. This is the actual thought behind this work. In this work, we achieve this objective by finding similar patterns by estimating pattern bounds. The maximum and minimum bounds of temporal patterns are found using novel expressions. We demonstrate a case study to show working of algorithm.