Bibliographic citations
Tasayco, C., (2021). Segmentación de usuarios que visitan el sitio web de una empresa utilizando la regresión logística con la técnica de sobremuestreo [Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/4945
Tasayco, C., Segmentación de usuarios que visitan el sitio web de una empresa utilizando la regresión logística con la técnica de sobremuestreo []. PE: Universidad Nacional Agraria La Molina; 2021. https://hdl.handle.net/20.500.12996/4945
@misc{renati/244552,
title = "Segmentación de usuarios que visitan el sitio web de una empresa utilizando la regresión logística con la técnica de sobremuestreo",
author = "Tasayco Silva, Carlos Marcial",
publisher = "Universidad Nacional Agraria La Molina",
year = "2021"
}
The monograph work was developed in a leading consulting company in Latin America specialized in analytical solutions for the digital marketing areas of multinational companies. The work consisted in the implementation and automation of a binomial logistic regression model to describe the segments of users who visit one of the client's website. Therefore, to perform the aforementioned statistical model, it started from the univariate analysis of each independent variable, followed by a “SMOTE” oversampling technique to avoid the imbalance of classes in the dependent variable, then with a matrix confusion with an accurate of 76% until finally validating the predictability by analyzing the ROC curve. The results of the research helped to demonstrate and conclude that there are variables that are significant in determining whether a user visiting the Web makes a transaction. Users who use organic channels without advertising media as a reference and direct or paid such as social networks negatively contribute to the likelihood of the user making the transaction. It was also observed that both the time of the visit or if the user visits the Web recurrently contribute to the probability of making the transaction is increased. Finally, the segmentation was performed based on the scores calculated by the binomial logistic regression to have three welldifferentiated segments that are high, medium and low probability. At the end of the work, the company accepted and showed its satisfaction with the results obtained
This item is licensed under a Creative Commons License