Open abuzarmahmood opened 8 months ago
Use additional shorter windows to differentiate spikes which have different "elbows"
Can also used Neighborhood component analysis or LDA to find what time-points in the spike-snapshot are MOST useful for separating waveforms.
Weighted PCA using the below weights might be useful
Using PCA of logistic regression coefficients (weighted transform) as transform improves classification accuracy
Use additional shorter windows to differentiate spikes which have different "elbows"
Can also used Neighborhood component analysis or LDA to find what time-points in the spike-snapshot are MOST useful for separating waveforms.