Skip to main content
. 2018 May 2;17(1):27–42. doi: 10.1007/s12021-018-9375-z

Fig. 4.

Fig. 4

Results of toolbox validation analyses using simulated data. (a) Classification accuracies for four separate analyses, each classifying between one noise dataset (consisting of independent samples of Gaussian noise, mean = 0, SD = 1) and one signal dataset consisting of Gaussian noise plus a signal. This signal consisted of values 0.05, 0.1, 0.2 or 0.3 added to the first 10 channels during timepoints 51–100. Accuracy scores were averaged across 10 cross-validation steps and 10 analysis repetitions. (b) Absolute SVM feature weights during time window 51-60 ms, averaged over cross-validation steps and analysis repetitions. Larger feature weights are visible for channels 1–10 in datasets with larger signals relative to the noise