期刊名称：Bonfring International Journal of Data Mining
印刷版ISSN：2250-107X
电子版ISSN：2277-5048
出版年度：2016
卷号：6
期号：2
页码：10-15
DOI：10.9756/BIJDM.8134
语种：English
出版社：Bonfring
摘要：Single Nucleotide Polymorphisms(SNP) are the foremost common type of genetic variation in human comprising nearly1/1,000th of the typical human genome. SNP offer the foremost complete information for genome-wide association studies. We tend to propose a process methodology to quickly notice true SNPs in public-available leukemia cancer database. Much analysis has been specializing in various genetic models to spot genes that may predict the disease status. However, increasing the amount of SNPs generates large amount of combined genetic outcomes to be tested. Classification could be a data processing technique to predict cluster membership for data instances. The ACO could be a probabilistic technique for computational issues which may be reduced to finding sensible ways through graphs. In this research paper data mining classification techniques linear classifier are analyzed with ACO on leukemia cancer dataset. Performance of these techniques is compared by accuracy, Sensitivity and Specificity. The experimental results show that Linear classifier with ACO is able to distinguish cancer diseases from normal with the maximum accuracy of 73.20%, Sensitivity of 69.21% and specificity of 65% whereas SVM are 70.00% of accuracy, 65.20% of sensitivity and specificity of 63.53%.
关键词：SNP; Single Nucleotide Polymorphism; Data Mining; ACO; Classification; Linear Classifier
