Contrary to Process Mining, it already exists a famous methodology to analyze and modelize data: CRISP-DM. This methodology stands for “Cross-Industry Standard Process for Data Mining” and is a widely-used methodology for data mining and predictive modeling. It was created by a group of data mining experts in the mid-1990s and was developed as a standard approach to data mining that could be applied across different industries, hence the name “Cross-Industry Standard Process for Data Mining” (CRISP-DM).
The group consisted of representatives from SPSS, Teradata, NCR, SAS Institute, SGI, IBM and other companies. The group’s goal was to create a standardized process for data mining that would be easy to understand and use by practitioners in different industries. The first version of the methodology was released in 1996. Since then, it has become one of the most widely-used data mining methodologies in industry and academia. However the purpose of Data Mining is quite different from Process Mining, so they have natural divergence, but also similarities:
The CRISP-DM process includes six stages of phases:
- Business Understanding
- Data Understanding
- Data Preparation
- Modeling
- Evaluation
- Deployment
Each phase includes specific tasks and deliverables to guide the data mining process from start to finish. The goal of CRISP-DM is to ensure that the business objectives are met in an efficient and effective manner.
In a way the ExYPro approach is similar to the CRISP-DM as it highlights clearly all the different steps to manage in an efficient way the Process exploration.
We could try to make some bridges between these two approaches however they have different purpose and they apply on different assets (data vs process):
