Sequence mining is one of the most investigated tasks in data mining and it has been studied under several perspectives. With the rise of Big Data technologies, the perspective of efficiency becomes prominent especially when mining massive sequences. In this paper, we perform a thorough experimental evaluation of several algorithms for sequential pattern mining and we provide an analysis of the results focusing on the different algorithmic choices and how these affect the performance of each algorithm. Experiments performed on real-world and synthetic datasets highlight relevant differences between existing algorithms and provide indications for Big Data scenarios.
An empirical evaluation of sequential pattern mining algorithms
Loglisci C.;Ceci M.;Malerba D.
2018-01-01
Abstract
Sequence mining is one of the most investigated tasks in data mining and it has been studied under several perspectives. With the rise of Big Data technologies, the perspective of efficiency becomes prominent especially when mining massive sequences. In this paper, we perform a thorough experimental evaluation of several algorithms for sequential pattern mining and we provide an analysis of the results focusing on the different algorithmic choices and how these affect the performance of each algorithm. Experiments performed on real-world and synthetic datasets highlight relevant differences between existing algorithms and provide indications for Big Data scenarios.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
