Association rule for purchase dependence in multi-item inventory

Abstract

In a multi-item inventory with very large number of items, purchase dependence amongst the items is often observed. In purchase dependence, stock-out of one item may result in non-purchase of another item which is in stock as purchase of the latter is dependent on the purchase of the former item. The primary objective of this research is to investigate whether association rules can be used to depict purchase dependence in retail sale. The investigation has been made with a case by analysing the impact of considering and not considering the association rules in inventory replenishment. Association rule correlates one item set with another item set and such rules can be mined from sale transaction data. Association rules have been mined from sale transaction data collected from the grocery stores. These rules have been considered for inventory replenishment. The impact on profitability for considering and not considering an association rule while replenishing the inventory has been analysed. Findings of the research show that total relevant cost of inventory comes down if the purchase dependence amongst the items depicted by association rule is taken into account while replenishing the inventory. Hence, association rules mined from transaction data can be used to depict purchase dependences which result in loss of profit if not addressed while replenishing inventory.