Predicción de clientes efectivos en la gestión de carteras de cobranza castigada en la empresa InteliBPO S.A.S a través de modelos de aprendizaje automático

Abstract

Generating value with data is a crucial point in any organization to stand out from the competition and continue to innovate, therefore, this degree project aims to make use of supervised machine learning algorithms focused on classification such as K-NN, vector support machines, random forests and decision trees, with the purpose of predicting the clients that will be effective in the collection management to be carried out by the InteliBPO S.A.S organization for portfolios in the penalty stage, based on the data recorded in the third quarter of the year 2022. A series of processes are carried out that include the exploratory analysis of the data, selection of the most representative attributes of the clients that add value to the models, a training stage and finally an analysis of the results obtained with the purpose of selecting the model. that it is more consistent with the prediction of effective records and that it can contribute positively to the generation of more assertive management strategies, presenting itself as a management support tool carried out by the organization's operations area.