<![CDATA[The advancement of information technology drives businesses to strategize to remain competitive. Retail businesses face challenges from changing consumer behaviors that prioritize convenience and speed in shopping, potentially reducing revenue if businesses fail to adapt quickly. An effective approach involves reanalyzing sales transaction data to identify consumer purchasing patterns, providing guidance for strategic decision-making. One effective technique is data mining using Market Basket Analysis models to analyze shopping baskets and identify correlations between items purchased. This model utilizes algorithms like Apriori and FP-Growth to generate association rules. Preprocessing the dataset to derive frequent itemsets, followed by applying association rules, helps identify significant correlations or patterns within the dataset. The Apriori and FP-Growth algorithms are applied with predefined minimum support and minimum confidence levels. Interpretation involves testing and verifying the discovered patterns against previous facts or hypotheses. The application of these algorithms shows their impact on the dataset size, as the chosen minimum support and confidence levels affect the number of association rules generated. Experiments with FP-Growth and Apriori algorithms indicate that using a minimum support of 0.02 requires longer execution time compared to 0.06. Using a minimum support of 0.02 and minimum confidence of 0.1 yields similar rules across different data divisions (30%, 40%, 50%, and 55% splits) of datasets containing 1200 and 2364 records out of 100% data. Both algorithms achieve 100% accuracy, demonstrating reliability and validity in discovering significant patterns within the same dataset]]>
Copyrights © 2024
