Shapley-based Structural Analysis of Neural Calibration for Stochastic Volatility Models
Abstract
Neural network-based approaches have emerged as efficient alternatives to traditional optimization-based procedures for the calibration of stochastic volatility models. However, existing work has focused primarily on predictive accuracy, with comparatively little attention devoted to understanding the structure of the learned inverse calibration mappings. In this work, we analyze neural calibration mappings for the Heston and rough Heston models across multilayer perceptron, highway, and softmax-parametrized highway architectures, using complementary Shapley-based methods from explainable AI. Specifically, we consider SHAP and νSHAP explanations, which capture distinct, complementary notions of feature relevance, corresponding to sensitivity and sufficiency of feature subsets, respectively. Short maturities and smile wings consistently dominate parameter inference, and the dominant attribution structure remains qualitatively stable across architectures despite differences in predictive accuracy and parameter count. Parameter-specific differences between SHAP and νSHAP further reveal how distinct regions of the implied volatility surface contribute to parameter recovery and expose substantial redundancy in the calibration input. Building on this redundancy, we show that νSHAP explanations can guide a significant reduction in input dimensionality for the rough Heston model while matching calibration accuracy relative to the full implied volatility surface. These findings demonstrate that complementary Shapley-based methods provide structural insight into learned inverse calibration mappings beyond predictive error metrics, and offer a practical route to feature selection in neural calibration problems.
Related
- Stochastic Knothe-Rosenblatt: Light-speed Calibration of Stochastic Local Volatility Models
- Structural limit of volatility target under stochastic volatility and discrete corrections
- Variance and volatility swap valuation with nonlinear liquidity-feedback stochastic volatility and regime switching
- On the Pricing of American Options under Stochastic Local Volatility and Stochastic Correlation via the RBSDE Framework
- A Design Concept of Forecasting Software for Normalized Vector Autoregressions with Fat Tails and Stochastic Volatility
BibTeX
@misc{afzali2026shapleybased,
title = {{Shapley-based Structural Analysis of Neural Calibration for Stochastic Volatility Models}},
author = {Shaïn Afzali and Serena Della Corte and Antonis Papapantoleon},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.03076},
eprint = {2610.03076},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.03076},
}
RIS
TY - UNPB
TI - Shapley-based Structural Analysis of Neural Calibration for Stochastic Volatility Models
AU - Shaïn Afzali
AU - Serena Della Corte
AU - Antonis Papapantoleon
PY - 2026
DA - 2026/10/02
PB - arXiv
DO - 10.48550/arxiv.2610.03076
UR - https://arxiv.org/abs/2610.03076
AB - Neural network-based approaches have emerged as efficient alternatives to traditional optimization-based procedures for the calibration of stochastic volatility models. However, existing work has focused primarily on predictive accuracy, with comparatively little attention devoted to understanding the structure of the learned inverse calibration mappings. In this work, we analyze neural calibration mappings for the Heston and rough Heston models across multilayer perceptron, highway, and softmax-parametrized highway architectures, using complementary Shapley-based methods from explainable AI. Specifically, we consider SHAP and νSHAP explanations, which capture distinct, complementary notions of feature relevance, corresponding to sensitivity and sufficiency of feature subsets, respectively. Short maturities and smile wings consistently dominate parameter inference, and the dominant attribution structure remains qualitatively stable across architectures despite differences in predictive accuracy and parameter count. Parameter-specific differences between SHAP and νSHAP further reveal how distinct regions of the implied volatility surface contribute to parameter recovery and expose substantial redundancy in the calibration input. Building on this redundancy, we show that νSHAP explanations can guide a significant reduction in input dimensionality for the rough Heston model while matching calibration accuracy relative to the full implied volatility surface. These findings demonstrate that complementary Shapley-based methods provide structural insight into learned inverse calibration mappings beyond predictive error metrics, and offer a practical route to feature selection in neural calibration problems.
ER -