Methodology for detecting the feasibility of using data mining in an organization

Abstract

This paper proposes a methodology to identify the feasibility of applying Data Mining techniques (DM) in an organization (institutions or enterprises). This methodology can be applied when there are different sources of data and knowledge, and their interactions are not well identified or defined. The methodology consists of five phases defined to know and characterize the enterprise processes, its its relationships, its actors, and sources of knowledge and data. As a final result, the methodology defines the problems of the institution/company that can be improved with DM tasks MD. The methodology selects the processes of interest based on prioritization, and identifies DM tasks to perform, from the present and expected future scenarios. The utilization of this proposition is illustrated in two different types of organization, health and petroleum.