Sequential pattern mining algorithms and their applications: a technical review

Sequential pattern mining (SPM) is a useful tool for extracting implicit and meaningful rules from sequence datasets that can aid the decision-making process. These rules are ordered pairs of events associated with their degree of occurrence and a user-defined support threshold. However, mining such rules from large datasets involves several challenges, such as data extraction and transformations, data partitioning and distribution, feature selection, processing capacity and environment, I/O overhead planning, and database scanning and pruning techniques. To address these challenges, several SPM algorithms and frameworks have been designed over the last few years that have been implemented in diverse domains and multidimensional datasets under variable environmental setups. However, there are still several challenges and limitations of the algorithms that need to be addressed for drawing the roadmaps for the future scopes of improvements. SPM has several implications in different fields such as market-basket analysis, treatment and diagnosis analysis, prediction purposes, bioinformatics, drug discovery, product quality testing, multimedia document planning, behavioural analysis, educational planning, recommender systems, business identification and diversifications, pollution control, sentiment analysis, IoT applications, tourism planning, and trajectory data analysis. In this article, a brief review on the premier sequential pattern mining algorithms developed over the last three decades is presented with a discussion on various recent application domains of the algorithms. A discussion on the limitations and challenges of the existing SPM algorithms with a viewpoint to address the open research dimensions is also appended.