Sistema de apoyo a la toma de decisiones aplicando Inteligencia de Negocio para identificar tendencias en la moda de vestir de empresas Retail moda.

Abstract

This project approach the problem faced by the Marketing department in a retail fashion company, where identifying the trends in dress fashion is a slow and imprecise process. The identification of the trends in dress fashion is a fundamental task in the process of marketing the products of a retail fashion company, used in the launch of a new product, special discounts, campaign promotions, etc. Proper marketing contributes to the increase in profitability of the company. To solve this problem, a decision support system is built applying Business Intelligence. The CRISP-DM methodology was used as a guide for software development; a Microsoft SQL Server database infrastructure with its Integration Services and Analysis Services components was implemented to exploit the data mining tools. The End User Interface is built in Power BI Desktop to display the results of the data analysis. Several prediction models were studied during the investigation: neural networks, decision tree and Bayes naive. Finally, it was determined that the most efficient model was Bayes naive, being able to handle classifications with multiple categories and also showed the highest percentage probability compared to the other models when examining the model elevation graph. This work can be used as a guide for other researchers who are dedicated to the study of Business Intelligence applied to Decision Support Systems.