Content area
Full text
ABSTRACT
Data mining techniques have led over various methods to gain knowledge from vast amount of data. So different research tools and techniques like classification algorithm, decision tree, association rules etc are available for bulk amount of data. Association rules are mainly used in mining transaction data to find interesting relationship between attribute values and also it is a main topic of data mining There is a a great challenge in candidate generation for large data with low support threshold. Through this paper we are making a study to show how association rules will be effective with the dense data and low support threshold. The data set which we have used in this paper is real time data of certain area and we are applying the data set in association rules to predict the chance of disease hit in that area using A Priori Algorithm. In this paper three different sets of rules are generated with the dataset and applied the apriori algorithm with it. With the algorithm, found the relation between the parameters in the database.
KEYWORDS: APriori algorithm, Association rules, Data mining, item based partitioning, multi Dimensional analysis.
(ProQuest: ... denotes formulae omitted.)
I. INTRODUCTION
Association rules discover correlation among data items in a transactional data base. It involves the discovery of rules that satisfy defined threshold from tabular database. Here the rule how often it occurs in the data base which is known as its frequency is important. Association rule mining is the process of finding frequent set with minimum support and confidence. The first phase is support counting phase where we have to find the frequent set generation. Effective partitioning may help for this process
We also have to create a border set to avoid frequent updating of real time data. In real life applications, the number of frequent sets are large in number and as the result, the number of association rules are also very large.. We are selecting only the rules which we have interested for disease prediction in this context. The discovery of frequent item sets with item constraints is also very much important.
There are many data mining algorithms such A priori Algorithm, Partition algorithm, Pincer-Search Algorith, Dynamic Itemset Counting Algorithm [2], FP-Tree Growth etc are...