Socioeconomic Shifts in Peru Before and After COVID-19: A Fuzzy C-Means Clustering Analysis

Abstract

Artificial intelligence, in its eagerness to mimic human reasoning, has incorporated fuzzy logic and fuzzy set theory, which allow modeling imprecise concepts and offer a flexible alternative to classical binary logic. The Fuzzy C-Means algorithm is presented as a fuzzy extension of traditional clustering methods, such as k-means, allowing a piece of data to belong to several clusters with different degrees of membership, better reflecting the ambiguity of many phenomena, such as human behavior. In the social sciences, it is common to look for patterns in the behavior of individuals using clustering techniques. However, these methods are often insufficient in the face of social complexity, where people may share traits from multiple categories simultaneously. This study characterizes the Peruvian population before (2018–2019) and after (2021–2022) the pandemic using data from the National Household Survey (ENAHO) and applies the Fuzzy C-means clustering algorithm to identify nuanced population profiles. The analysis reveals three distinct groups: one marked by persistent informality and vulnerability, another initially characterized by higher education and formal employment but increasingly precarious after the pandemic, and a third intermediate cluster reflecting social mobility and instability. Results show significant post-pandemic shifts, including increased formal employment in the transitional group and a reconfiguration of self-reported ethnic identities, especially among Quechua populations according to economic stratification. The study underscores the pandemic’s role in both reinforcing longstanding inequalities and generating new patterns of social and economic stratification.