A Hybrid Genetic-Relative Reduct Algorithm for Pre-Processing the Diabetic Dataset

Karamath Ateeq, Dr. Gopinath Ganapathy

Abstract


Diabetes is a high sugar problem that occurs because of the inadequate secretion of Insulin in the human body. Nowadays, data are accumulated in digital form. To extract the required knowledge from the data, Data Mining is suggested to be the best tool. Since, the accumulated data contains a lot of noisy, irrelevant and redundant data; the dataset should be pre-processed before extracting required knowledge. In this paper, a hybrid algorithm combining Genetic Algorithm and Relative Reduct Algorithm from Rough Set Theory is proposed. This algorithm is proposed to remove noisy and unwanted data. The proposed Hybrid Genetic-Relative Reduct Algorithm is compared with existing algorithms. The proposed Hybrid Genetic-Relative Reduct Pre-Processing Algorithm has reduced the number of data attributes to minimum. The number of reduced attributes and time taken for execution of the algorithm is taken to evaluate the performance of the algorithm. The results obtained support the proposed hybrid Genetic-Relative Reduct algorithm as the best pre-processor than the existing algorithms

Keywords: Pre-processing, Diabetes, Hybrid Algorithm, Genetic Algorithm, Rough Set Theory, Relative Reduct Algorithm.

Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v8i1.2853

Refbacks

  • There are currently no refbacks.




Copyright (c) 2017 International Journal of Advanced Research in Computer Science