Citas bibligráficas
Guillén, A., (2019). Diseño de un modelo de Data Mining para incrementar la disponibilidad de entrega de documentos en una empresa financiera [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/21296
Guillén, A., Diseño de un modelo de Data Mining para incrementar la disponibilidad de entrega de documentos en una empresa financiera [Tesis]. PE: Universidad Nacional de Ingeniería; 2019. http://hdl.handle.net/20.500.14076/21296
@misc{renati/711151,
title = "Diseño de un modelo de Data Mining para incrementar la disponibilidad de entrega de documentos en una empresa financiera",
author = "Guillén Vásquez, Aníbal Javier",
publisher = "Universidad Nacional de Ingeniería",
year = "2019"
}
This thesis has as one of its objectives the identification and classification of the documents that must remain in the In House Archive at the time of the transfer process, so that the indicator of Availability of the In House Archive increases above 90 %. The documents that must remain in the In House Archive are those most likely to be requested by users. Such identification is carried out with the use of Data Mining classification algorithms, which are a type of Data Mining problem that catalogs a set of data, usually with a binary value, in order to predict its behavior; This value is known as the Target variable. The CRISP-DM methodology is used, which is a road map that allows, through 6 stages, to cover the Data Mining cycle and obtain the expected result. These stages are: Understanding the business, Understanding the data, Preparing the data, Modeling, Evaluation and Deployment. It is decided to evaluate 4 classification algorithms: Decision trees, Naive Bayes, Rule induction and Logistic regression, and it is chosen that algorithm that has the highest Accuracy indicator and the lowest Specificity indicator, in this case the algorithm of Decision trees Implementing the model with the chosen algorithm, it is verified that the indicator of Availability of the In House Archive is increased to 92%, above the target set of 90%.
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons