Quality and Risk Management in Data Mining: A CRISP-DM Perspective.

Abstract

The area of data science knowledge responsible for dealing with this new reality is diffuse, including mathematics, statistics, computing, engineering, psychology, and administration, among many other areas that make up a new scenario that is still changing. Different models have emerged over the years to systematize the procedures to be followed. Among them, CRISP-DM (Cross Industry Standard Process for Data Mining) has become one of the most widespread in the industry. However, the lack of detailed instructions means the framework is often incorrectly used. Therefore, this research aims to present a utilitarian and didactic model based on the latest advances in the literature and through the lens of production engineering. In order to achieve this objective, exploratory research was carried out based on a systematic review and subsequent categorization of each of the CRISP-DM steps, detailing the authors’ contributions to each stage. In addition, it is proposed that guidelines from the areas of Quality Management and Risk Management be added to the subject, consolidating a useful and didactic model of relevance.