weekend37 / string-kernels

Implementation of the Polynomial String Kernel and linear time String Kernel implementation. Compatible and ready for easy plug-in to Scikit-Learn's Support Vector Machines and supports multithreading.

String Kernels

This package contains an implementation of the Polynomial String Kernel and a linear time String Kernel algorithm as described in our paper, High Resolution Ancestry Deconvolution for Next Generation Genomic Data.

It offers

• Linear time computation of two effective string kernels.

• Compatibility with Scikit-Learn's Support Vector Machines (easy plug-in).

• Multithreading.

Usage

To install the package, execute from the command line

pip install string-kernels

And then you're all set!

Assuming you have Scikit-Learn already installed, you can use Lodhi's string kernel via

from sklearn import svm
from stringkernels.kernels import string_kernel
model = svm.SVC(kernel=string_kernel())

and the polynomial string kernel,

from sklearn import svm
from stringkernels.kernels import polynomial_string_kernel
model = svm.SVC(kernel=polynomial_string_kernel())

For morer information read the docs or take a look at the notebook example.ipynb for further demonstration of usage.

If you end up using this in your research we kindly ask you to cite us! :)