Optimal Inventory Classification using Data Mining Techniques

Introduction

Inventory control is the process whereby the investment in materials and parts carried in stocks is regulated, in order to keep the total cost associated with the system to a minimum. This requires familiarity with supply sources, modes of transportation, budgeting, physical handling, price negotiations including bulk quantity discounts, record keeping, and monitoring the incoming quality of items. In real life situations, we have a very large number of items in an inventory and it is not computationally feasible to set stock and service control guidelines for each individual item. ABC classification is usually applied to control the inventory level of the items. ABC classification is based on Pareto’s Law, which states that a small percentage of items accounts for a large percentage of value. This value can be sales, profits, or other measure of importance. Roughly 10 percent to 20 percent of inventory items account for 70 percent to 80 percent of inventory value. These highly valuable items are classified as A inventory items. Moderate value items account for approximately 30 percent of inventory items and contribute to roughly 35 percent of the total. They are called B items. Finally, approximately 50 percent of the items only contribute to roughly 10 percent of total inventory value. These are called C items and are of least importance. An ABC classification needs not to be done only on the basis of the dollar usages, because for some items the value not only comes from the item’s own dollar usage, but also from its influence on the dollar usages of other items. Therefore, such influences related to different items should be considered in classification of inventories. Classifying inventory based on degree of importance allows us to give priority to important inventory items and manage those with care.

For many items, however, ABC classification is not suitable for inventory control. Managers have to shift some items among categories for a number of reasons. Several researchers considered there may be other criteria that represent important considerations for management. The certainty of supply, the rate of obsolescence, and the impact of a stock out of the item are all examples of such considerations. Several researchers suggested that multiple criteria should be used in the classification of inventories (Flores &Whybark 1987; Cohen & Ernst 1988; Lenard & Roy 1995). However, the problem is that the profit of one item not only comes from its own sales, but also from its influence on the sales of other items or reverse, i.e., the ‘cross-selling effect (Anand et al., 1997).In such a situation, it should be explained clearly whether the cross-selling effects would influence the ranking of items or not, and how to group the items if such effects existed, not concerning what and how many criteria could be used. In this chapter, we have discussed three cases for classification of inventories with cross-selling effect.

Firstly, the cross-selling with lost profit of item/item-set can be defined as a criterion for evaluating the importance of item. The loss profit of item is the total profit that the item may takes away when it is out of stock. The results indicate that a considerable large part of inventory items change their positions in the ranking list of traditional ABC classification. Many items that traditionally do not belong to A group have been moved into the A group by the cross-selling effect, and also many items that traditionally belong to C group have been promoted into higher group because of their high loss profits.

Secondly, the cross-selling with loss profit can be extended by including similarity among transactions. Clustering algorithm clusters transactions that contain similar items. Under this classification scheme, the clustering algorithm is used to partition the transactional database into different clusters. Apriori algorithm is utilized for mining association rules from each cluster to find frequent items. Later the loss profit is calculated for each frequent item. The obtained loss profit is used to rank frequent items in each cluster. Some items that traditionally do not belong to the A group in each cluster have been moved into the group A by the cross-selling effect to reconfigure their inventory policies, and also some items that traditionally belong to C group in each cluster have been promoted into higher group because of their high values of loss profits and should not be ignored as these were treated before. Thus, the ranking of frequent items in each cluster facilitate optimal inventory control.