Skip to main content
. 2016 Jun 14;7:157. doi: 10.3389/fphar.2016.00157

Figure 9.

Figure 9

SVM classification accuracy increases with more parameters then decreases due to “curse of dimensionality”—sparseness of parameter vectors relative to dimension. Best classification accuracy from all combinations of y parameters (params) using bottom/top SPI firing rate percentiles on x-axis.