A Handbook of Mathematical Models with Python by Dr. Ranja Sarker

Author:Dr. Ranja Sarker
Language: eng
Format: epub, mobi
Publisher: Packt Publishing Pvt Ltd
Published: 2023-08-22T00:00:00+00:00

Figure 5.5b: Sigmoid kernel (L), data classified using sigmoid (R)

All SVM kernels have a parameter that trades off the misclassification of the dataset against the simplicity of the separator. While training an SVM with the RBF kernel, which is an exponential () function, the parameter a is greater than zero and defines the influence of a training example on the separator. The selection of the parameters in respective kernel functions is critical to an SVM’s performance.

In imbalanced datasets, the parameters dedicated to providing weights on classes and samples become significant, as they might be required to give more importance to a certain sample or class in such cases. The effect of sample weighting on the class boundary is shown in Figure 5.6, wherein the data point size is proportional to the sample weights:

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.