Publication

Abstract : Fragmentation in a distributed database is a design technique that reduces query processing time by keeping the relation size small. When it comes to storing high-dimensional data in a distributed manner, the processing time increases. This is due to the huge attribute size. In this paper, a method is proposed which can reduce the size of high-dimensional data by using feature selection technique. This technique reduces dimensions by removing irrelevant or correlated attributes from the dataset without removing any relevant data. The algorithm used for feature selection and vertical fragmentation is the random forest and Bond Energy Algorithm (BEA), respectively. Experiments show that our method can produce better fragments.