Study of Some Properties of the K-Mode Clustering Method Through Resampling, Applied to the Making Business Decisions in the Framework of Covid in Uruguay

EasyChair Preprint no. 8986

Abstract

In a study in the field of corporate finance, developed during the pandemic in Uruguay, it is inquired about the COVID-19 IMPACT ON THE URUGUAYAN COMPANIES in decision-making on downsizing, telecommuting, training and work environment among others. For that, it resorts to the construction of a typology of business decision making using the method- do of clustering kmodes of modal
profile on variables of binary type. Are analyzed several scenarios that account for the number of decision-making profiles based on the number of groups. This document presents the study of the performance of a clustering method, the modal
profile kmodes, taking into account account the variability of some metrics that allow evaluating the homogeneity of the groups and decide on a possible solution to the number of groups. that variability the method, which is random in the choice of center, was studied based on the starting others, thus seeing the dependency with the solution found and also in the structure of the data for which we work with
learning samples, varying its size by resampling. It works with data that comes from a study in the field of corporate finance, developed during the pandemic In Uruguay.

Keyphrases: Clustering, K-modes, remuestreo

@Booklet{EasyChair:8986,
  author = {Diego Araujo and Ramón Álvarez-Vaz and Mauro De la Vega},
  title = {Study of Some Properties of the K-Mode Clustering Method Through Resampling, Applied to the Making Business Decisions in the Framework of Covid in Uruguay},
  howpublished = {EasyChair Preprint no. 8986},

  year = {EasyChair, 2022}}