Network Science for Time Series Clustering and its Applications

Clustering in case of time series datasets has been an active area of research since decades. The accumulation of continuous data points introduces new patterns, that cannot be captured easily when a large time series dataset is considered. Although there exist several pairwise time series distance computation measures, they cannot implicitly capture the underlying data patterns. Also, with the increased data size, distance computation comes at the cost of time and space complexity. Therefore, in this paper, we discuss how network science can help in clustering, that facilitate in study and analysis of repeating patterns, atypical patterns and also capture patterns with changing seasons in case of multiple time series domain.