Table 1 Performance of AFs generated for a four-class music genre problem [46].
From: Analytical Features: A Knowledge-Based Approach to Audio Feature Generation
| Â | Â | Perf train | Perf test | Feature |
|---|---|---|---|---|
Best features | REF | 58% | 53% | Variance(Bandwidth(Envelope(FFT(Split(x, 250)), 50), 10)) |
| Â | EDS | 61% | 64% | Log(Range(SpectralSpread(Hann(Split(Blackman(x), 153))))) |
Best classifiers | REF | 79% | 80% | Multilayer perceptron (23 features, 3 layers) |
| Â | EDS | 82% | 80% | kNN (25 features, 8 neighbors) |