Sequential pattern mining using IDLists

Sequential pattern mining is a practical problem whose objective is to discover helpful informative patterns in a stored database such as market transaction databases. It covers many applications in different areas. Recently, a study that improved the runtime for mining patterns was proposed. It was called pseudo-IDLists and it helps prevent duplicate data from replicating during the mining process. However, the idea only works for the special type of sequential patterns, which are clickstream patterns. Direct applying the idea for sequential pattern mining is not feasible. Hence, we proposed adaptions and changes to the novel idea and proposed SUI (Sequential pattern mining Using IDList), a sequential pattern mining algorithm based on pseudo-IDLists. Via experiments on three test databases, we show that SUI is efficient and effective regarding runtime and memory consumption. © 2020, Springer Nature Switzerland AG