Proceso para el descubrimiento de conocimiento en las bases de datos de la Universidad de Cuenca mediante técnicas de Data Mining

Abstract

Considering the large volume of data that the University has collected for several decades in its information systems, the present work proposes the use of this valuable resource through the definition of a specific process for the development of Data Mining projects within of the University of Cuenca; which, based on the formal industry standard (CRISP-DM), proposes an agile methodology approach, and seeks to use the University's Data Warehouse system as the main source for data access. This established process has been validated with its application to three specific problems: a) identification of the parameters that are most related to academic success in new students, b) construction of a model to predict student dropout in new students and c) group analysis on the relationship between student qualification and teacher evaluation. As a result, it was confirmed that the application of these techniques are a decisive contribution to Data-Driven Decision Making. On the other hand, it was verified the results obtained in each problem have generated a new knowledge in each case, besides that the process has been defined as an important aid for the organized development of the project, generating sufficient documentation for its replication and later analysis.