SVM

By syuoni (https://github.com/syuoni)

Introduction

A implementation of Support Vector Machine (SVM) with Python.

Since a deeply vectorized impelementation of Sequential Minimal Optimization (SMO) algrithm, the computation is quite fast. An alternative optimization method of SLSQP (from scipy) is provided.

Demonstration

For example,

>>> from SVM import SVM, VecKernel
>>> svm = SVM(SVM(kernel_gram=VecKernel.gaussian())
>>> predictor = svm.construct_predictor(y, X, method='smo')
>>> predictor.predict(x_test)