| Autors: Ivanova, M. S., Tchobanova, Z. N. Title: The Influence of Regression Kernel Functions at Predictive Modelling in Electronics Keywords: machine learning, regression analysis, support vector machin Abstract: The paper presents an approach for predicting the power dissipation in electronic circuits through utilization of support vector machine, which is an algorithm from controlled machine learning. Three predictive models are created, outlining the power dissipation at different values of power supply and load resistance. The models are evaluated considering four regression kernel functions: dot, polynomial, radial and epachnenikov. The results show that the most suitable kernel function can be found and used for creation of a concrete model with a small error. References Issue
Copyright University of Niš, Faculty of Electronic Engineering |
Вид: пленарен доклад в международен форум, публикация в реферирано издание
