Análisis predictivo en la segmentación de usuarios

Abstract

Abstract: This curricular integration work is focused on performing predictive analysis with machine learning techniques for user segmentation using an E-commerce database. The main objective of the curricular integration work is to apply Machine Learning models to predict customer preferences and behaviors. This analysis aims to develop new marketing and user loyalty strategies with companies. For this analysis, the CRISP-DM methodology was used, which allowed us to carry out an exhaustive analysis of the selected data set, performing cleaning, transformation and modeling to reach the results. Unsupervised Machine Learning algorithms such as the K-means and DBSCAN algorithms were also implemented for segmentation, and the supervised Random Forest or Decision Tree algorithm was implemented for prediction. The results obtained demonstrated the effectiveness of machine learning techniques that allowed identifying patterns and creating homogeneous groups, which allows improving business decision making and personalizing offers. This study provides key tools to maximize customer values in E-commerce database environments.