Implementación de machine learning para clasificación de clientes en la empresa VALENTINA EXPORT S.A.

Abstract

This thesis details the application of logistic regression as an innovative solution to segment the customers of Valentina Export S.A. The methodology used involves the transformation of variables and data cleaning to ensure the quality and accuracy of the model. Logistic regression is presented as a key tool to predict the probability of a customer being “export eligible” or “export ineligible” based on various parameters such as purchase history, financial liquidity, and order size. The practical implementation of this model not only helped optimize product allocation to strategic customers, but also created more effective strategies for decision-making and personalized customer service. This machine learning approach aims not only to improve the operational efficiency of Valentina Export S.A., but also to provide informed insight for decision-making in a dynamic and challenging demand environment.