Sporathe knowledge network
ES
Universidad EAFIT

Maestría en Ingeniería · 2025

Aproximación para personalización de interfaces basado en datos históricos en un caso aplicado de la banca

This project proposes an approach to generate personalization proposals through a user grouping model based on pattern identification in transaction history, with the aim of improving each user's experience by allowing them to view options that suit their particular needs. The project has two stages: the development of the user grouping model using machine learning (ML) algorithms through the CRISP-DM methodology, achieving a result of six user clusters with a silhouette coefficient of 64%, which represents adequate cohesion and distinguishable patterns; and secondly, the development of personalized proposals using the design thinking methodology. This was the result of a case study applied to a digital channel of a financial institution in Colombia, where, based on customer validations, a 90% favorability rating was obtained in terms of the perceived value of the proposal made for the specific cluster of each group, with an average of 30% more compared to the control variants.

Full text 88 pages with text

Read the full thesis Record in the repository

Contents

Keywords

Its library's thesaurus