Identifying Possible Socio-Demographic Factors in Infant Mortality Rate Classification With Orthogonal Projections to Latent Structures Discriminant Analysis

¹ Department of Mathematics, Gadjah Mada University, Yogyakarta, Indonesia
² Department of Statistics, Institut Sains & Teknologi Akprind Yogyakarta, Indonesia

^* e-mail: novianapratiwi@mail.ugm.ac.id

Abstract

The measurement of infant mortality rate (IMR) stands as a crucial and complex indicator within global public health. Particularly for countries with constrained resources necessitating easily calculable and accurate gauges of population health, IMR retains its relevance. Several studies investigating infant mortality rates adopt a multilevel perspective and connect underlying factors with proximate variables. Conversely, interdisciplinary viewpoints, such as socio-demographic, offer a comprehensive overview of the intricate connections between socio-demographic disparities and mortality rates. This research focuses on the analysis of infant mortality rates, aiming to identify the sociodemographic factors influencing them to facilitate the formulation of more targeted intervention strategies. Employing Orthogonal Projection to Latent Structures - Discriminant Analysis (OPLS-DA) as the analytical method is motivated by its capability to navigate complex multivariable structures and elucidate the relationships between socio-demographic variables and infant mortality rates. The OPLS-DA model is used to extract significant patterns and relationships in the data so that it is possible for socio-demographic factors to have a significant impact on infant mortality rates. It is hoped that the results of the classification will provide insight and become the basis for developing policies by considering socio-demographic aspects in efforts to prevent infant mortality.