S. Vijayarani and R. Prasannalakshmi
Bharathiar University, India

Data mining technology is engaged in establishing helpful and unfamiliar data from the huge databases.Generally, data mining methods are useful for static databases for knowledge extraction wherever currently available data mining techniques are not appropriate and it also has a number of limitations for managing dynamic databases. A data stream manages dynamic data sets and it has become one of the essential research domains in data mining. The fundamental definition of the data stream is an arrival of continuous and unlimited data which may not be stored fully because it needs more storage capacity. In order to perform data analysis with this, many new data mining techniques are to be required. Data analysis is carried out by using clustering, classification, frequent item set mining and association rule generation. Association rule mining is one of the significant research problems in the data stream which helps to find out the relationship between the data items in the transactional databases. This research work concentrated on how the traditional algorithms are used for generating association rules in data streams. The algorithms used in this work are Assoc Outliers, Frequent Item sets and Supervised Association Rule. A number of rules generated by an algorithm and execution time are considered as the performance factors. Experimental results give that Frequent Item set algorithm efficiency is better than Assoc Outliers and Supervised Association Rule Algorithms. This implementation work is executed in the Tanagra data mining tool.

Data Stream, Association Rules, Assoc Outliers, Frequent Item sets and Supervised Association Rule,Tanagra.