Looking into the possibility of novel dissimilarity measure to discover similarity profiled temporal association patterns in IoT

Abstract

The main objective is to address the possibility of designing a similarity measure and using the same to find temporal similarity between patterns in temporal databases which are time stamped. The proposed similarity measure is a function of threshold and deviation. Given, a reference pattern which is of user interest, unearthing eccentric temporal association patterns requires an efficient similarity or correlation measure which can estimate the association pattern similarity degree. This research studies the similarity measure for revealing analogous or similar temporal patterns by using the designed similarity measure. A case study is demonstrated to show working of the proposed measure.