Association Pattern Analysis of Global Company Market Capitalization Using the FP-Growth Algorithm with Load Balancing Constraint

Abstract

This research focuses on analyzing the global company market capitalization dataset using the FP-Growth algorithm combined with a load-balancing constraint approach. The main objective is to identify association patterns among different market capitalization categories Small, Medium, Large, Mega, and Ultra to understand their distribution and interrelationships. The study begins with data preprocessing, cleaning, and categorization of companies based on their market values. The FP-Growth algorithm is applied with a minimum support threshold of 0.02, and a load balancing constraint is introduced by filtering rules with support ≥ 0.05 and lift > 1, ensuring balanced and significant association patterns. The analysis results show that the most dominant categories are Medium and Small, representing the majority of companies worldwide, while Large, Mega, and Ultra categories are relatively rare. The strongest rule indicates that countries with “Large” companies are very likely to also have “Small” and “Medium” companies. Evaluation metrics show an average lift of 1.171 and an average confidence of 1.000, confirming strong and reliable associations. Overall, this study provides insights into global market capitalization patterns and demonstrates the effectiveness of FP-Growth with constraints in revealing meaningful, balanced relationships within large-scale business data.   Keywords – FP-Growth, Load Balancing Constraint, Market Capitalization, Association.