Analogy of the Application of Clustering and K-Means Techniques for the Approximation of Values of Human Development Indicators

Abstract

The objective of this study was to apply Clustering and K-Means' techniques to classify the departments of Peru according to their Human Development Index. In this article, the elbow method was used to determine the optimal number of clusters, applying the classification algorithms to group the departments of Peru according to their similarities, in addition to the Principal Component Analysis (PCA) technique for a better display of clusters. After applying the unsupervised algorithms, the results were more relevant in clusters 2 and 4 according to their HDI, made up of the departments of Arequipa, the Constitutional Province of Callao, Ica, Lima, Moquegua and Tacna, where the most notable is the life expectancy at birth, the population with full secondary education, the number of years of education, the average per capita income, and the state's density index. The results obtained by the K-Means algorithm show more cohesive results than the Clustering algorithm.