Indian Journal of Science and Technology
DOI: 10.17485/ijst/2016/v9i28/87995
Year: 2016, Volume: 9, Issue: 28, Pages: 1-6
Original Article
Navneet* and Nasib Singh Gill
Department of Computer Science and Applications, [email protected]
[email protected]
*Author for correspondence
Navneet
Department of Computer Science and Applications,
Email: [email protected]
Objectives: To optimize data mining technique for big data. Methods/Analysis: This paper designs a technique for the data mining of big data by modifying the existing data mining technique using fuzzy and neural network. The present technique firstly performs the dimension reduction. Then reduced dimension datasets are clustered, while the remaining attributes are used to classify such dataset by using automated fuzzy. Findings: The existing data mining techniques are not optimized on such data. The simulation using the fuzzy on various dataset shows the optimization of technique. The RNFCA algorithm is analyzed adding the RNFCA algorithm to the WEKA library on the Intel i5 @ 2.67 GHz using the eclipse IDE. The algorithm is analyzed on the datasets having 400 instances with 25 attributes and 32561 instances with 15 attributes. The detail description of these datasets is given in table 2. The performance of the RNFCA algorithm can be compared with existing CCSA algorithm and the decision tree i.e. J48. The figure 4 -7 shows the comparison graph of the J48, CCSA and RNFCA over various parameters. Applications/Improvement: The simulation using the fuzzy on various dataset shows the optimization of technique.
Keywords: BIG Data, Dimension Reduction, Fuzzy, K-Mean, Neural, Schwarz Criteria
Subscribe now for latest articles and news.