Association Rule Mining Algorithms for Brain Tumour Detection

Abstract: In biomedical field there is lots of data regarding patient reports, as there is daily nearly 10,000 brain tumor patient MRS image is used to find out patient suffering from which tumor type. It is difficult to predict the tumor type with the use of metabolite value like CHO, CR, CR2, and NAA for large database. Association Rule mining is one of the fundamental research topics in data mining and knowledge discovery that identifies interesting relationships between item sets and predicted the associative and correlative behaviour for new data. So author used three association rule algorithms: Apriori Association Rule, Predictive Apriori algorithm and Filtered Associator algorithm to use best rules for finding out the tumor type. These three algorithms presented at different support and confidence level, author used weka 3.7.7 data mining tool to get best rules for finding tumor type. It was found that the entire three algorithms give the best rules to easily find out the tumor type for large database.

Keywords: fMRI, Association Rule mining, Apriori algorithm, Predictive Apriori, Filtered Associator, Confidence level, Support level.