Progression analysis of signals: Extending CRISP-DM to stream analytics

Stream analytics focuses on analysis of signals generated simultaneously and over time. The specific patterns in the signals can indicate some of the outcomes such as failure of a device, etc. Therefore, novel ways to find specific patterns in the signals generated by many sources are required. In this paper, we extend the CRISP-DM process to include data preparation approaches for sequence mining. We present progression analysis, an approach for converting streams of records to be able to detect useful signals for analysis. To illustrate the process, we present a healthcare example where patients diagnosed with Tobacco Use Disorder develop multiple other diseases over multiple hospital visits. The common sequences of the diseases diagnosed in the TUD patients over multiple hospital visits are presented and discussed. Finally, the generalizability of the progression analysis is discussed.