Large-scale online learning of implied volatilities

作者：

Highlights：

• A network for estimating the Black–Scholes implied volatilities is proposed.

• Large-scale online learning is employed to eliminate generalization errors.

• An iterative method further improves the estimates of the network.

• Incredible high accuracy is achieved over a surprisingly wide input domain.

• By utilizing the GPU, the time for the computation significantly is reduced.

摘要

•A network for estimating the Black–Scholes implied volatilities is proposed.•Large-scale online learning is employed to eliminate generalization errors.•An iterative method further improves the estimates of the network.•Incredible high accuracy is achieved over a surprisingly wide input domain.•By utilizing the GPU, the time for the computation significantly is reduced.

论文关键词：Implied volatility,Black–Scholes model,Online learning,Large-scale data,Deep learning

论文评审过程：Received 9 July 2021, Revised 18 November 2021, Accepted 25 April 2022, Available online 5 May 2022, Version of Record 13 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117365