MARKET BASKET ANALYSIS USING ASSOCIATION RULE MINING

Abstract

Market Basket Analysis (MBA) is a widely used data mining technique that identifies relationships between products purchased together in transactional datasets. With the rapid expansion of retail and e-commerce industries, massive volumes of transactional data are generated daily, making it essential for businesses to adopt intelligent analytical tools that can extract meaningful patterns from this data. Traditional MBA techniques generally rely on offline batch processing methods that analyze static datasets and fail to respond quickly to dynamic changes in consumer purchasing behavior. This limitation reduces their effectiveness in modern retail environments where purchasing trends evolve continuously. The proposed system introduces an Intelligent RealTime Market Basket Analytics Platform that enhances classical association rule mining by incorporating automated model updates, real-time analytics, and user-friendly visualization capabilities. The system processes transactional data obtained through manual entry and CSV-based uploads and stores it in a relational database that acts as a centralized data repository. Machine learning models are automatically retrained whenever new data is added to ensure that generated patterns accurately reflect current purchasing trends. The core analytical component utilizes the FP-Growth algorithm to efficiently discover frequent itemsets while optionally comparing performance with the Apriori algorithm. Association rules are evaluated using support, confidence, and lift metrics to identify strong product relationships and generate actionable insights. The proposed platform enables retailers to optimize product placement, improve cross-selling strategies, and enhance marketing campaigns. By integrating real-time learning with interpretable rule generation, the system provides a scalable and intelligent decision-support tool for modern retail analytics