Skip to main content
. Author manuscript; available in PMC: 2013 Jun 15.
Published in final edited form as: Nanotechnology. 2012 Jun 15;23(23):235101. doi: 10.1088/0957-4484/23/23/235101

Table 4.

The top nine parameter combinations together with the SVM setting.

Cumulative Accuracy (%) SVM setting Parameters Used
84 Unsealed ClusterOnTime(%) clusterfreq3 clusterfreq8 clusterfreq9
80.8 Unsealed Spike Amplitude (pA) NumPeakslnCluster ClusterOnTime(%) freq3 clusterfreq1 clusterfreq2 clusterfreq3 clusterfreq5 clusterfreq6 clusterfreq7 clusterfreq8 wavelet5 wavelet7
80.7 Easy NumPeakslnCluster ClusterOnTime(%) freq2 clusterfreq2 clusterfreq7 clusterfreq8 wavelet2
80.7 Easy NumPeakslnCluster clusterfreq1 clusterfreq4 clusterfreq5 clusterfreq7 waveletl wavelet3 wavelet7
80.7 Unsealed Spike Amplitude (pA) Spike Frequeney (spikes per 4000 samples) NumPeakslnCluster ClusterOnTime(%) clusterfreq2 clusterfreq3 clusterfreq4 clusterfreq5 clusterfreq7 wavelet2 wavelet4 wavelet7 wavelet9
80.5 Unsealed NumPeakslnCluster ClusterOnTime(%) freq2 clusterfreq2 clusterfreq7 clusterfreq8 wavelet2
80.5 Easy Spike Amplitude (pA) NumPeakslnCluster ClusterOnTime(%) freq3 clusterfreq1 clusterfreq2 clusterfreq3 clusterfreq5 clusterfreq6 clusterfreq7 clusterfreq8 wavelet5 wavelet7
80.3 Unsealed Spike Amplitude (pA) NumPeakslnCluster ClusterOnTime(%) freq1 freq4 clusterfreq1 clusterfreq2 clusterfreq5 clusterfreq7 clusterfreq8 clusterfreq9 wavelet2 wavelet3 wavelet5
80.1 Easy Spike Width (Samples) ClusterOnTime(%) freq1 freq2 freq4 freq7 clusterfreq2 clusterfreq4 clusterfreq5 clusterfreq6 clusterfreq7 waveletl wavelet2 wavelet3 wavelet4 wavelet6

The SVM settings are as follows: Easy: Easy.py is a predefined python script that is distributed with LIBSVM to automatically determine a few of the adjustable parameters of the SVM. The script iteratively searches the SVM parameters (gamma, C) to specify the most accurate kernel. Scaled: Before training, both the training and testing datasets are scaled so all the parameters range from -1 to 1. This helps to prevent one parameter from overwhelming the SVM data. Unsealed: The SVM is trained with data that has not been scaled.