Hands-On Artificial Intelligence for IoT

Classifying wine using SVM

We'll use the svm.SVC function provided by the scikit library for the task. The reason to do so is that the TensorFlow library provides us, as of the time of writing, with only a linear implementation of SVM, and it works only for binary classification. We can make our own SVM using the maths we learned in previously in TensorFlow, and SVM_TensorFlow.ipynb in the GitHub repository contains the implementation in TensorFlow. The following code can be found in the Wine_quality_using_SVM.ipynb.

The SVC classifier of scikit is a support vector classifier. It can also handle multiclass support using a one-versus-one scheme. Some of the optional parameters of the method are as follows:

• C: It's a parameter specifying ...