Elaboración de una propuesta tecnológica basada en algoritmos de aprendizaje automático para fidelización de clientes.

Abstract

Machine learning is present in different modern industries, but it is popularly applied in the commercial sector for customer analysis, its application allows to discover patterns in customer behavior that companies can use to apply business strategies, such as retaining or building customer loyalty. Grouping or clustering is a machine learning technique for this type of analysis, it is based on the partition of a data set into several groups where each group contains elements like each other and maintains a difference with respect to the other groups. The objective of this titling work is to obtain the segmentation of the clients of the National Lottery company through the application of automatic learning algorithms, for which variables were created that allowed identifying the level of loyalty of the clients of the National Lottery company. For the development of this degree work, the CRISP-DM methodology was applied, which served for the data mining process. Data analysis was performed based on the RFM model (Recency, Frequency, Amount) and the k-means, k-nearest neighbor, and decision tree clustering algorithms were applied to this model. To validate the result of the algorithms, the data was separated for training and tests that allowed evaluating the accuracy of the algorithms, finally the Power BI tool was used to present the results in a friendly and simple way.