Association Rule Mining (ARM) requires only transaction data, meaning it does not use labeled data or a target variable like in supervised learning. The dataset must be structured in a one-hot encoded format, where each row represents a transaction (or instance), and each column represents an item (or attribute) with binary values (1 or 0). A value of 1 indicates the presence of an item in that transaction, while 0 indicates its absence. This format allows ARM algorithms, such as the Apriori algorithm, to identify patterns and associations between different items based on their co-occurrence across multiple transactions. Unlike classification or regression models, ARM focuses on discovering relationships within unlabeled data rather than predicting an outcome.