Data Mining and Knowledge Engineering

E. Ramaraj ¹, N. Venkatesan

Affiliations
1 Madurai Kamaraj University, Madurai, Tamilnadu, IN

Abstract

The tremendous growth in data has generated the need for new techniques that can intelligently transform the massive data into useful information and knowledge. Data Mining is such a technique that extracts non-trivial, implicit, previously unknown and potentially useful information from the data in databases. Association Rule Mining is one of the most important and well-researched techniques of data mining. It aims to extract interesting correlations, frequent patterns, associations of casual structures among sets of items in the transaction databases or other data repositories. Association rules are widely used in market databases, spatial databases, biological databases, medical databases and crime databases. This paper focuses a new algorithm to mine both positive and negative rules from the real-time surveyed medical database. Association rules are defined as implication of the form A->B where A and B are frequent itemsets in a transaction database. This new algorithm extends this definition to include association rules of forms A ->^B, ^A -> B and ^A -> ^B, which indicate negative associations between itemsets is called negative rules. Negative rules are generated from infrequent itemsets using multidimensional data model.

Keywords

Data Mining, Association Rules, Infrequent Itemsets Negative Rules.

Full Text

     

E. Ramaraj ¹, K. Ramesh Kumar ²

Affiliations
1 Computer Centre, Alagappa University, Karaikudi-630002, Tamilnadu, IN
2 Department of Computer Science and Engineering, Alagappa University, Karaikudi-630002, Tamilnadu, IN

Abstract

Association rule mining is to extract the interesting correlation and relation between the large volumes of databases. Association rule mining process is divided into two sub problem: The first problem is to find the frequent itemsets from the transaction and second problem is to construct the rule from the mined frequent itemset. Frequent itemsets generation is the prerequisite and most time overwhelming process for association rule mining. Apriori algorithm is the familiar and fundamental algorithm to generate the frequent itemsets from the transaction sets. Till now, Lot of researcher modified the Apriori in various manner like partition approach, Hash function and etc. But most efficient Apriori-like algorithms rely heavily on the minimum support constraints to prune the vast amount of non-candidate itemsets. These algorithms store many unwanted itemsets and transactions. In this paper propose a novel frequent itemsets generation algorithm. The drawback of the HEA, AprioriTId and Apriori overcome by the proposed algorithm. The proposed algorithm is an improved version of High Efficient AprioriTid (HEA) algorithm. The proposed algorithm is using the two theorems which are proposed in this paper. The proposed algorithm is tested with the synthetic retail dataset. It performed well at low supports. The experimental reports also show that proposed algorithm on an outset is faster than HEA, AprioriTID and Apriori.

Keywords

Data Mining, Association Rule Mining, Frequent Itemsets, Transaction Reduction.

Full Text