A Naïve Approach to High Utility Rare Itemset Mining Algorithm using Temporal Concept – THURI

Abstract: Business strategies use information about any organizations past performance that can be used to predict its future performance. The right business strategy can be formulated by clearly understanding the dynamically changing business environment. Temporal data mining can be used for obtaining temporal measures of various operations of company. Temporal mining can be instrumental for tracking the temporal changes in the business activities which allows the company for gaining insights to process improvement and optimization. An appealing question in temporal mining that is concerned with business strategy is to study the contribution of past decisions on the success of the organization. Temporal Data Mining is an important step in the Knowledge Discovery process that discovers temporal patterns or temporal rules from Temporal Databases. Temporal Data Mining Algorithms are the algorithms which consider temporal patterns from temporal data or fit models to temporal databases. Traditional Frequent itemset mining discovers frequent itemsets from transactional databases using only items occurrence frequency and not considering items utility. But in many real world situations, utility of itemsets based upon user’s perspective such as cost, profit or revenue is of significant importance.
One of the latest data mining research areas is Utility Mining which emphasis on all types of utility factors and incorporates utility concepts in data mining tasks. The utility-based descriptive data mining which aims at discovering itemsets having high total utility is termed as High utility itemset mining. High Utility itemsets may contain frequent as well as rare itemsets.Temporal data mining is a very fast expanding field with many new research results reported and many new temporal data mining analysis methods or prototypes developed recently. The temporal significant rare utility itemsets are those itemsets which appear infrequently in the current time window of large databases but are highly associated with specific data. In this paper, a novel method is proposed, namely THURI (High Utility Rare Itemset Mining Algorithm using Temporal Concept), for efficiently and effectively mine high utility rare itemsets from databases with temporal consideration of utility values. The novel contribution of THURI is that it can effectively extract high utility rare itemsets from temporal transaction databases.

Keywords: Frequent Itemset Mining, Rare Itemset Mining, Utility Mining, High utility Rare Itemset Mining, Temporal Mining