Aplicación del aprendizaje automático en el análisis de cartera de una empresa pública de agua , alcantarillado y aseo

Abstract

Currently, new trends and technological innovations such as the application of machine learning are of great importance for companies and industries to promote efficiency in decision making. In this sense, the Public Water, Sewerage and Cleaning Company (EPAAA) of the Pasaje canton, seeks to support its decisions based on the portfolio analysis for the company's collection management, in view of this problem, a Decision Support System was developed (DSS) applying business intelligence, data mining and machine learning, using the CRISP-DM methodology which consists of the phases: business understanding, data understanding, data preparation, modeling, evaluation and deployment. With the implementation of the DSS it is possible to determine what is happening in terms of the portfolio and collection management; In addition, it is possible to characterize the clients, classify the clients by sector and cycle, classification of paying clients, classification of debtor clients and support for portfolio management. The DSS is composed of a dashboard interface consisting of statistical graphs, classifications and subscriber predictions. The results obtained with the implementation of the DSS in the EPAAA allowed a 94.44% efficiency in decision-making, a result that was obtained through a survey of the collaborators who make decisions in the EPAAA. In conclusion, this work allows us to improve the response times of the collaborators, generating an efficient management of collections, the DSS managed to strengthen the knowledge about the portfolio, allowing to characterize the clients, an important achievement is that the clients could be classified according to to the debt where as a result of this, better decisions could be made.