Table 1 Hyper-parameters most appropriate values (in samples).
From: Speech emotion recognition via graph-based representations
Information | (Samples) | EMODB | DEMoS | AESDD |
|---|---|---|---|---|
Statistical | Window | 10000 | 8000 | 8000 |
Overlap | 2000 | 6000 | 6000 | |
Structural | Window | 200 | 100 | 200 |
Overlap | 50 | 50 | 100 |
