A Practical Guide to SVM with TensorFlow

This code belongs to the "A Practical Guide to SVM" blog post.

We implement the pratical procedure proposed by Chih-Wei Hsu, Chih-Chung Chang, and Chih-Jen Lin in their paper A Practical Guide to Support Vector Classification.

Support Vector Machine (SVM) is a popular technique in machine learning for classification and regression. It is highly preferred by many as it produces significant accuracy with less computation power. The paper proposed a simple procedure to obtain reasonable results from SVM.

Dataset

In this work, we use the Astroparticle dataset which is a binary classification dataset. A preprocessed version of the dataset can be downloaded here. The dataset contains 4 features and is splitted into training set and test set. The training set contains 3089 rows and the the test set contains 4000 rows.

Model performance

The following tableau summarize the performance of the models we have implemented with the proposed procedure:

	Accuracy	Precision	Recall	F1-Score
Simple SVM	0.95875	0.95875	0.95875	0.95874
SVM with RBF kernel	0.96725	0.96726	0.96725	0.96724

We see that we have some improvement in all the metric with the RBF kernel.

Useful Links

• http://danielhnyk.cz/creating-your-own-estimator-scikit-learn/
• https://scikit-learn.org/dev/developers/contributing.html#rolling-your-own-estimator
• https://github.com/nfmcclure/tensorflow_cookbook/tree/master/04_Support_Vector_Machines/04_Working_with_Kernels