On Calibration Neural Networks for extracting implied information from American options

Bibliographic record. Follow the original-source link for the publication.

Field	Value
Primary domain	Volatility
Other domains	—
Methods	Financial Ml
Facets	—
Authors	Shuaiqiang Liu, Álvaro Leitao, Anastasia Borovykh, Cornelis W. Oosterlee
Published	2020-01-31
Source	arXiv Quantitative Finance History
Identifiers	arxiv:2001.11786
URL	Open original source

Editorial synthesis

Why it matters

CaNN replaces repeated online inversion for American-option implied parameters with an offline neural approximation, targeting the computational bottleneck in extracting volatility and dividend yield. (abstract:S2, abstract:S5, abstract:S6)

Main author claims

• The authors report: A machine learning approach can estimate Black-Scholes implied volatility and dividend yield for American options in a fast and robust manner. (abstract:S2)
• The authors report: Approximating the inverse function with a neural network decouples offline training from online prediction and eliminates the need for iterative online processes. (abstract:S3)
• The authors report: The introduced Calibration Neural Network (CaNN) framework can simultaneously estimate multiple parameters. (abstract:S5)

Data, method, or discussion scope

The study proposes a data-driven method that uses neural networks to approximate the inverse functions for implied volatility and dividend yield on a specified computational domain, with the CaNN framework for multi-parameter calibration. No specific empirical data or benchmark comparisons are provided in the abstract. (abstract:S3, abstract:S5)

Main limitations

The method relies on the Black-Scholes model assumptions and may not generalize directly to other pricing models. The abstract does not report performance on real market data or accuracy comparisons with traditional methods. (abstract:S2)

Relationships

• None recorded.