A method and system for data mining is disclosed in which the premise and/or conclusion of an association rule can include overall attributes of an entire basket (e.g., its total dollar value). The items contained in a single aggregate sale (e.g., all of the purchased items in a particular market basket, referred to herein as a "market basket grouping") are characterized according to predetermined attributes. Each attribute is identified and an "imaginary item" is included in the data for each market basket grouping which possesses an identified attribute. When the data is subjected to traditional association analysis, the imaginary items are included in the analysis and may be utilized to identify frequent itemsets that are typically found in market basket groupings having the identified characteristics.