<![CDATA[The rapid growth of e-commerce has driven an increase in household product transactions, including furniture, resulting in large-scale and diverse transactional data. However, most furniture transaction datasets exhibit sparse characteristics, as each order typically contains only one or a small number of items, making it difficult to generate frequent itemset combinations in association rule analysis. This study applies the Apriori algorithm to analyze online furniture sales transaction data from 2025, consisting of 1,938 records and 14 attributes. The research stages include data cleaning, transaction transformation, one-hot encoding, and the determination of a minimum support threshold of 0.0005 and a minimum confidence threshold of 0.1, adjusted to the characteristics of the dataset. The results indicate the formation of 421 frequent itemsets that meet the specified criteria, with a dominance of single-item itemsets caused by the low variation of items within each transaction. The analysis further shows that the generation of association rules is highly limited and fails to produce meaningful product relationship patterns. This limitation is primarily attributed to the highly sparse nature of the dataset, with an average of one item per transaction. This study demonstrates that applying the Apriori algorithm to highly sparse datasets results in very limited patterns and is dominated by single-item itemsets. These findings serve as an empirical study and a cautionary tale regarding the limitations of the Apriori algorithm when applied to transaction data with an average of one item per transaction, particularly in generating longer itemsets. Overall, this research contributes as an empirical warning on the limitations of applying the Apriori algorithm to highly sparse transaction datasets and emphasizes the importance of analyzing data characteristics prior to implementing association rule mining on e-commerce platforms. Â ]]>
