یک روش نظری ارزش چند معیاره برای رتبه بندی قوانین انجمنی برای استفاده از تصمیم گیری بهتر کسب و کار

Market Basket Analysis or Association Rule Mining (ARM) is an approach to discover the purchase patterns of the customers by extracting and analyzing the basket of items which sell together. Businesses are also keen to discover those rules which can generate more profits. These profitable purchase patterns of customers once identified can lead to better product assortment decision making for businesses. Better Product Assortment decisions can surely be a competitive advantage for businesses in terms of customer satisfaction and profit generation. The concepts of support and confidence in the association rules help to extract the rules with frequent and reliable co-occurrences of the items in customers’ purchases. The profitable rules can be assessed using the domain-related measures such as item set value and the cross-selling profit associated with the association rules. We propose a ranking mechanism to combine the different criteria of Confidence, Support, Item Set Value and Cross-Selling Profit to get an overall interestingness measure; using Multi-Attribute Value Theory (MAVT) approach; which in turn uses DIVIZ as the implementation tool. These association rules can be used as a leverage for the marketing activities like cross-selling promotions, shelf placement etc. and other crucial decisions like product assortment selection.