sdjkjsdh / adaptive_kernel_methods

Python code for Kernel LMS (KLMS) algorithm with sklearn API. We also implemented neural spike kernels. We also have KLMS and Kernel PCA for neural spike data.

Adaptive Kernel Methods following scikit-learn API

Here we mainly test the Kernel LMS algorithm for R^n vectors and spike trains. KernelLMS deals with Rˆn vectors and SpikeKLMS with spike times series To test those methods either run 'test_KLMS.py' or 'test_SpikeKLMS.py'. This will show how the methods work for regression. There also an example of KLMS for classification at 'test_klms_classify.py'

Spike_KLMS algorithm used SLASH library for spike train signal processing, which is included here. SLASH provides spike-spike, population-spike and population-population inner products. To test SLASH for unserpervised learning, we included Paiva's PCA at 'test_paiva.py' and a modification for populations at 'test_populationPCA.py'