Table 11 The Bates parameters are estimated with Bates–CaNN, by calibrating to a data set (35 samples) generated by the Bates model. In DE, the random seed is 2 and the population size is \(10\times N_{v}=80\)
From: A neural network-based framework for financial model calibration
Parameters | CaNN search space | Bates | Calibrated |
|---|---|---|---|
Intensity of jumps, \(\lambda _{J}\) | [0,3.0] | 1.0 | 1.065 |
Mean of jumps, \(\mu _{J}\) | [0,0.4] | 0.1 | 0.087 |
Variance of jumps, \(\nu ^{2}_{J}\) | [0,0.3] | 0.160 | 0.146 |
Correlation, ρ | [−0.9,0.0] | −0.3 | −0.228 |
Reversion speed, κ | [0.1,3.0] | 1.0 | 0.598 |
Long average variance, ν̄ | [0.01,0.5] | 0.1 | 0.128 |
Volatility of volatility, γ | [0.01,0.8] | 0.7 | 0.776 |
Initial variance, \(\nu _{0}\) | [0.01,0.5] | 0.1 | 0.102 |
Total Squared Error | – | – | 4.95 × 10−6 |
Function evaluation | – | – | 842,800 |
Time (seconds) | – | – | 1.8 |