A New Partition Based Association Rule Mining Algorithm for BigData

P. Prabakaran, Dr.K.Ramesh Kumar


Association Rule Mining is an important research area in the field of Data Mining especially in case of ‘Sales transactions’. A number of algorithms have been presented in this regard. In this paper a comparison of PARTITION algorithm with CMA algorithm is presented after improving the PARTITION algorithm. In this study, randomized partitioning of database is done. The database is randomized so that real random data is available for better results. The randomized partitioning of database has been implemented in different tool, i.e., VB. Net, as compared to CMA, which uses MATLAB for randomization so as to achieve better performance and efficient results. In the end it has been proved with extensive experiments that although Randomized PARTITION algorithm takes two database scans as compared to CMA that takes single database scan, still it gives better results with more efficiency than CMA.

Keywords: Data Mining, Association Rule Mining, Randomized PARTITION Algorithm

Full Text:


DOI: https://doi.org/10.26483/ijarcs.v6i6.2525


  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science