| publication name | The impact of scaling on Support Vector Machine in Breast Cancer Diagnosis |
|---|---|
| Authors | Elsayed Badr, Mustafa Abdulsalam, Hagar Ahmed |
| year | 2020 |
| keywords | |
| journal | |
| volume | Not Available |
| issue | Not Available |
| pages | Not Available |
| publisher | Not Available |
| Local/International | International |
| Paper Link | https://www.researchgate.net/profile/El-Sayed-Badr/publication/345208854_The_impact_of_scaling_on_Support_Vector_Machine_in_Breast_Cancer_Diagnosis/links/5fa4453ea6fdcc062415ae4b/The-impact-of-scaling-on-Support-Vector-Machine-in-Breast-Cancer-Diagnosis.pdf |
| Full paper | download |
| Supplementary materials | Not Available |
Abstract
By using support vector machine (SVM) and the grid technique Badr et al.[1] introduced new scaling techniques on the data set Wisconsin from UCI machine learning with a total 569 rows and 33 columns. These scaling techniques overcame the standard normalization techniques. In this paper, three new scaling techniques are proposed by using SVM and the grid technique on the the data set Wisconsin from UCI machine learning with a total 569 rows and 32 columns. These scaling techniques are:(i) de Buchet for p=(∞)(ii) Lp-norm for p=(∞)(iii) Entropy. Experimental results show that SVM with new scaling techniques achieves 98.60%, 98.42% and 98.42% accuracy against to the standard normalization by 96.49%.