Utility Itemset Mining Models using Multi Objective Utility Factors with Reduction in Search Space

Abstract

Discovering knowledge from the corpus of transactions is the crucial objective of the newlinedata mining. Default methods of data mining use confidence and their support of the newlineitems appear in transactions of the given corpus for scaling their priority and the newlineimportance of their associability. However, the support of the elements in the given newlinecorpus often claims is least significant for the discovery of knowledge from the given newlinecorpus of transactions such as market-basket data. This is so, due to the appearance of newlinethe confidence and support of the items in the transactions of the market-basket data newlinecorpus which are not substantial for the identification of the most profitable items and newlinepattern of these items. In this context, the utility of the item instead of frequency of newlineitem is coined for the discovery of the knowledge from the given corpus of data. newlineUtility indicates the factors of the data context such as profit to conclude the newlineimportance of the items and their association in the given corpus of transactions. The newlineusage of such factors for the discovery of knowledge from the given corpus is referred newlineto as utility mining. The high utility itemset mining is the buzz of the recent past newlineresearch contributions of the data mining. This research work endeavours to find fast newlineand siginificant utility mining models with the objective of reducing mining process newlinecomplexity using multi-objective utility factors. newlineThe initial phase of this research work portray the scope of the objectives through a newlinedetailed review of the research contributions observed in contemporary literature newlinerelating to utility-based mining models. A utility mining model has been devised and newlinemonitors the marginal profits from the sale of the pattern of items, and in the event of a newlinechange in providing a profit from the sale of an itemset. It changes the priority of the newlinecorresponding itemset based on the change observed in the profit that considered as newlineutility. The proposed algorithm works and generates results as similar as the newlineconventional algorithm. The propo