Article Preview
TopIntroduction
With the advent of information technology, data resources have become progressively increasing, however the information contained in data resources has not been completely investigated and used. To overcome this problem data mining techniques came into focus. Data mining refers to extracting useful information from vast amounts of data. Knowledge or insight discovered using data mining helps in more effective individual and group decision making. There are various data mining techniques like association rule mining, clustering, temporal association rule mining, classification, sequence mining, utility mining etc. Data mining had been used in a variety of applicative areas such as marketing, customer relationship management, engineering, and medicine analysis, expert prediction, web mining, mobile computing and inventory management.
Further, mining of association rules in extensive databases is a very much considered system in the field of data mining with strategies like Apriori (Agrawal et al., 1993; Agrawal & Srikant, 1994). In the association rule mining area, most of the research efforts went in the first phase to improve the algorithmic performance and in the second phase into reducing the output set by allowing the possibility to express constraints on the desired results (García et al., 2006). Over the last decade a diffusion of algorithms that deal with those troubles through the refinement of search strategies, pruning techniques and information systems have been developed.
The alignment of the data mining process and algorithms with the extensive economic objectives of the tasks supported by data mining is essential so as to permit the additional impact of data mining on business applications. The traditional association rule mining techniques consider the utility of the items via its presence in the transaction set. The frequency of item-set is not enough to reflect the real utility of an item-set. For example, the income supervisor may not be interested in common item-sets that do not generate enormous earnings. Recently, one of the maximum tough data mining responsibilities is the mining of high utility item-sets successfully. Identification of the item-sets with excessive utilities is called as utility mining. The utility can be measured in terms of cost, income or different expressions of user possibilities. For instance, a pc machine can be extra profitable than a telephone in terms of profit. The ultimate economic utility obtained as the outcome of the data mining product has the impact of all the diverse stages of the data mining processes.
Thus, utility mining is valuable in an extensive variety of practical applications and was as of late concentrated by (Chan et al., 2003; Liu & Qu, 2012; Tseng et al., 2006; Yao et al., 2004; Yao et al., 2006). The high utility item-set mining is to find all item-sets that have utility larger than a user specified value of minimum utility. Traditional high utility mining algorithms assume that items have only positive unit profit. However, in real-world, items may appear with negative unit profit also. For example, if a customer bought a smart TV then he would have received one pen-drive free of cost in a promotional scheme of a retail store. In this case the retail store loses the profit for each unit of pen-drive given away. Although this results in a loss, the profit from selling a smart TV is more advantageous to the store. This example explains the concept of mining negative unit profit of item values in utility mining. Many researchers have devoted their time to study about utility mining (Ahmed et al., 2009; Chan et al., 2003; Li et al., 2008; Liu et al., 2005; Yao et al., 2006), but none of them used it as decision making system for inventory management. By utility mining, several important business area decisions such as maximizing revenue, minimizing marketing or inventory costs can be considered.