Machine Learning Application Development Guidelines Using CRISP-DM and Scrum Concept
Muhammad Tamiramin Hayat Suhendar, Yani Widyani
Abstract
In the field of data mining, machine learning (ML) has been utilized in the search for solutions to various problems. One widely used model process for ML application development is the Cross Industry Standard Process for Data Mining (CRISP-DM). On the other hand, Scrum has emerged as the most popular agile method for software development in recent years. In this research, we proposed an ML application development guideline for data mining by incorporating relevant Scrum concepts into CRISP-DM. The process involves analyzing CRISP-DM and the development situation through interviews with experienced ML software developers. Furthermore, an analysis of the implementation of Scrum concepts in CRISP-DM is conducted. The proposed guideline is represented in Essence and tested through a case study, qualitative evaluation, and evidence map. The evidence map is used to analysis the importance of proposed guideline components is examined. The results indicate that the proposed guideline can be utilized to assist in the development of ML software.