So far, we are familiar with both waterfall-based and iteration-based development cycles. The rapid development of Artificial Intelligence (AI) and Machine Learning (ML) makes it quite difficult for vendors to follow the development methods used because AI, DM, and ML involve datasets. So the data and methods in ML cannot be separated.

One of the development processes currently used is the Cross Industry Standard Process for Data Mining (CRISP-DM). This process integrates DM modeling into the development process. Especially in the data understanding to evaluation section.

After CRISP-DM was used to create Data Mining-based applications, some developers needed a new process standard specifically for ML, especially due to the rapid development of Deep Learning. So that raises CRISP-ML where ML is slightly different from DM. An integration with quality assurance (QA) results in the CRISP-ML(Q) development model.

In accordance with its meaning, ML requires a learning process before inference, which is usually unsupervised. For more details, please see the following video.