Up

Performance evaluation of Apriori algorithm using association rule mining technique

File Size:
1.16 MB
Volume:
Volume 2, Issue 5 (May, 2016)
Publication No:
IJTC201605002
Author:
Jasleen Kaur, Rasbir Singh , Rupinder Kaur Gurm
Downloads:
14 x

Abstract
Many researchers studied the mining of data to get required information and knowledge in business applications, medical field, engineering projects and market data store. Discussing various data mining techniques and challenges posed by them. Association Rule mining proved to be best in many fields which generate relations between the items and generates strong rules using artificial intelligence with minimum confidence. These association rules is built into a large decision support system which can take managerial and operational decisions based on knowledge base developed and maintained by company. The most widely used Apriori algorithm for generating association rule discovers frequent patterns by candidate key generation which is a costly and memory consuming algorithm. Improvement in Apriori is ongoing topic these days. FP-Tree is another enhanced apriori algorithm which constructs frequent pattern tree without candidate generation. This makes it cost effective and more efficient. Other algorithms like K-means, Hash tree, bayesian networks, neural networks are discussed briefly. The proposed work based on making Apriori less complex by transposition techniques. Further improvement can be done by altering support count formula and reducing the number of transactions.

Keywords:

Data mining, Apriori algorithm, Association Rule Mining, FP-Tree algorithm, efficiency

Tags Associated: