Akademik Veri Yönetim Sistemi
International Conference on Artificial Intelligence and Applied Mathematics in Engineering (ICAIAME), Antalya, Türkiye, 20 - 22 Nisan 2019, ss.544-558
In recent years, Data Mining has grown significantly in almost every field. Sectors such as banking, insurance, pharmaceuticals and retailing utilize data mining techniques widely to reduce costs, improve research and increase sales. However, large projects are being carried out on this issue and standards on data mining technique are required. In order to respond to this request, it can be said that CRoss-Industry Standard Process for Data Mining (CRISP-DM) is the most important effort. CRISP-DM is used in many studies, grew as an industry standard, and is defined as a series of sequential steps that guide the application of data mining technique. The CRISP-DM reference model for data mining provides an overview of the life cycle of a data mining project and includes the phases, related tasks and outputs of a project. CRISP-DM is an effort to provide industrial standards for DM applications, including business understanding, data understanding, data preparation, modeling, evaluation and deployment steps. The quality and accuracy of each of the CRISP-DM steps related to DM applications used in different fields is very important for the success of the whole project. In this study, firstly the CRISP-DM algorithm and steps are investigated, later, the data related stages of the CRISP-DM usage in a project (data monitoring and evaluation) are examined, and explained by using an example application. In this process, the supplier database based on the CRISP-DM algorithm for Data Mining has been analyzed and processed. The aim of this study is to explain the role of the steps of CRISP-DM, and specially to process, understand and prepare information in the process of data discovery. In this way, it is expected that create awareness about CRISP-DM, and the impact of this process on projects is expected to be clearly understood.