AN INTELLIGENT REAL-TIME MARKET DAY IDENTIFICATION FRAMEWORK USING MACHINE LEARNING FOR SOCIOECONOMIC ANALYSIS OF TRADITIONAL MARKETS IN KOGI STATE

Abstract

Periodic markets play a critical role in rural socioeconomic systems. Yet, their operations are often constrained by limited access to accurate market-day information and a lack of data-driven insights into their economic impacts. This study presents an integrated framework for identifying real-time market days and modeling the socioeconomic impact of periodic markets in Kogi State, Nigeria. To realize this objective, an application was developed to digitally identify and disseminate market-day information for major markets operating in the state. This will greatly improve accessibility and planning for market participation. To measure the impact of such a system, machine learning and data analysis techniques were applied in modeling the socioeconomic viability of the day-to-day activities in the markets using market-related dataset with 53 instances and market-related indicators like population density, transportation cost variations, market location, establishment type, and environmental conditions. Decision Tree–based models were employed to capture nonlinear relationships and assess the relative importance of these factors in explaining commodity price reduction during market days. The results indicate that population density, market location, and market establishment type are the most influential determinants of price behavior, while environmental and institutional factors also play meaningful roles. The study demonstrates that integrating real-time market information systems with machine learning analytics provides actionable insights for traders, consumers, and policymakers. This approach contributes to improved market efficiency, informed decision-making, and sustainable planning of periodic markets in developing economies.