FAQ

What is JKernelMachines?

JKernelMachines is a java library for learning with kernels. It is primary designed to deal with custom kernels that are not easily found in standard libraries, such as kernels on structured data. Check the wiki for more information.

What is the targeted audience?

It is primarily designed for researchers looking for Kernel Machines with exotic kernels. If you are looking for an ultra-fast classifier to put into production, you would better look at other popular software. JKernelMachines is meant as a library providing learning algorithms compatible with any kernels, which means you will have to code your final program.

You provide the kernel, with care for the learning, handling data, cross-validation, etc.

Why should I use it?

It's free (as in free speech), very easy to use, reasonably fast, with a lot of features.

What do I need in order to use JKernelMachines?

A working jdk, and that's all! JKernelMachines is a completely standalone Java library, meaning that no additional jar files are needed. However, most algebra computation can be accelerated using the EJML library. To use it, just copy the EJML jar file in JKernelMachines folder before launching ant, and the script will automatically detect it and compile some wrappers. At run time, if the EJML jar is in the classpath, the wrappers are used for matrix operations.

Can I contribute?

Of course! Contributions are encouraged. If you feel a fundamental kernel is missing, you can reach the author (picard -at- ensea -dot- fr) for inclusion in future releases.