The goal of periodic pattern mining is to find some patterns that are repeating over time in a sequence of events or symbols. For example, a simple periodic pattern that can be found in customer data is that someone buys wine with cheese about every 7 days. This pattern is illustrated below:
Finding such periodic pattern can be useful. For instance, if we know that a customer has this habit, we can send him a message every week to remind him of buying wine and cheese or offering him some discount. Such pattern can also help to understand the behavior of customers, and thus be used for taking decisions.
Besides, shopping data, periodic pattern mining can be applied to many other applications. For example, periodic patterns could be found in a sequence of events observed in a computer network, or periodic patterns could be found in the locations visited by a person.
To apply periodic pattern mining, the SPMF software provides open-source efficient implementations of many algorithms and variations. These algorithms can be used to find periodic patterns in a single sequence or multiple sequences. The SPMF software can be downloaded from the website: http://www.philippe-fournier-viger.com/spmf/ .
To install the software, you may follow the instructions on the download page of that website. Then, you may check the documentation page which provides examples of how to run various algorithms such as PFPM and MPFPS for periodic pattern mining. Besides, you may check the datasetspage of that website provides several benchmark datasets for testing the algorithms and comparing their performance.