Table 5. Directed hypothesis testing for valence.
| Hypothesis | Estimate | Est. Error | CI Lower | CI Upper | Evid. Ratio | Star |
|---|---|---|---|---|---|---|
| Roughness < 0 | -0.08 | 0.01 | -Inf. | -0.06 | Inf. | * |
| Harmonicity > 0 | 0.06 | 0.01 | 0.05 | Inf. | Inf. | * |
| SpectralEntropy < 0 | -0.06 | 0.01 | -Inf. | -0.04 | Inf. | * |
| 12-TET Dissimilarity < 0 | -0.04 | 0.01 | -Inf. | -0.02 | Inf. | * |
| AveragePitch > 0 | 0.40 | 0.03 | 0.35 | Inf. | Inf. | * |
| AveragePitch^2 > 0 | 0.16 | 0.02 | 0.12 | Inf. | Inf. | * |
| GMSI:Roughness < 0 | -0.03 | 0.02 | -Inf. | 0.01 | 8.11 | |
| GMSI:Harmonicity > 0 | 0.03 | 0.01 | 0.00 | Inf. | 38.13 | * |
| GMSI:SpectralEntropy < 0 | 0.00 | 0.02 | -0.04 | Inf. | 0.68 | |
| GMSI:12-TET Dissimilarity > 0 | 0.00 | 0.01 | -Inf. | 0.02 | 1.34 | |
| GMSI:AveragePitch < 0 | -0.12 | 0.05 | -Inf. | -0.03 | 64.69 | * |
| GMSI:AveragePitch^2 > 0 | 0.02 | 0.01 | 0.00 | Inf. | Inf. | * |
* = The expected value under the hypothesis lies outside the 95%-CI (credibility intervals). Estimate = mean.
Estimate error = standard deviation of the posterior distribution. CI lower and CI upper are two-sided 95% credibility intervals. Evidence ratio = the posterior probability under the hypothesis against its alternative.