A comprehensive study on periodicity mining algorithms

Mining knowledge from time series database is always a challenging task due to inherent complexity. Periodicity mining is one of the methods for analysing time series data in order to find regularities and other significant statistics. It has been studied for variety of applications such as stock market analysis, customer purchase sequence analysis, electricity power consumption, weather forecasting, spatiotemporal, social networks, biological data analysis etc. It reveals important observations and inferences about the temporal behavior and trends from the data under study and would assist in more effective strategy planning. There exists a number of algorithms in literature for periodicity mining. This paper presents a comprehensive study of the research to date on periodic pattern mining algorithms. It gives brief categorization of different types of periodicity, the mining techniques used and its merits and demerits. It will act as a compendium on periodicity mining for interested researchers in identifying promising research direction for further exploration.