Bibliographic citations
Elguera, R., (2018). Segmentación de clientes de un casino utilizando el algoritmo partición alrededor de medoides (PAM) con datos mixtos [Tesis, Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/3312
Elguera, R., Segmentación de clientes de un casino utilizando el algoritmo partición alrededor de medoides (PAM) con datos mixtos [Tesis]. : Universidad Nacional Agraria La Molina; 2018. https://hdl.handle.net/20.500.12996/3312
@misc{renati/247062,
title = "Segmentación de clientes de un casino utilizando el algoritmo partición alrededor de medoides (PAM) con datos mixtos",
author = "Elguera Vega, Rhony Miguel",
publisher = "Universidad Nacional Agraria La Molina",
year = "2018"
}
Currently, the large amount of data stored by customers in different companies and the processing capacity provided by computers have generated great interest in research; as well as, develop methods and algorithms for grouping analysis. Clustering methods aimed at customer segmentation allow companies to identify patterns and profiles of purchase or services, helping them to make better decisions on channel and advertising strategies for their clients. In the present investigation the grouping method based on the partitions of k-Medoids with the PAM (Partition Around Medoids) algorithm is applied. The PAM algorithm is based on partitioning the data set into k groups, where k is known; is considered more robust to atypical data and noise, is based on minimizing the sum of dissimilarities between an object and the Medoid (center of the group). The objective of this research is to apply the PAM algorithm to segment the customers of a casino with the data obtained, through the use of cards in the slot machine. The silhouette method allowed to identify three clusters as the optimal number. The cluster analysis with the PAM algorithm using the Gower distance measure, resulted in the segmentation of clients for the three clusters with percentages of 49.4%, 11.3% and 39.4% respectively. The grouping was validated, obtaining all the significant ANVAs for the 6 quantitative variables and 99.35% accuracy with the C5.0 classification tree. The results of the characterization show that Cluster 1 are clients with averages values for the 6 variables in an intermediate level, 67.0% are men and 100% the card type is classic. In Cluster 2 there are the clients with the highest values in the average of the 6 variables, 59% are men and 100% use the silver card. In Cluster 3, customers with the lowest averages are found, 64% are men and 100% use classic cards
This item is licensed under a Creative Commons License