Volatility Forecasting with Machine Learning and Intraday Commonality¶
Bibliographic record. Follow the original-source link for the publication.
| Field | Value |
|---|---|
| Primary domain | Volatility |
| Other domains | — |
| Methods | Financial Ml, Research Methods |
| Facets | — |
| Authors | Chao Zhang, Yihuang Zhang, Mihai Cucuringu, Zhongmin Qian |
| Published | 2023-03-20 |
| Source | Journal of Financial Econometrics |
| Identifiers | doi:10.1093/jjfinec/nbad005 |
| URL | Open original source |
Editorial synthesis¶
Why it matters¶
This peer-reviewed version shares the same core claims as the earlier preprint and thus is structurally more usable as a benchmark; still, universality and superiority claims remain boundary-sensitive. (abstract:S1, abstract:S2, abstract:S3, abstract:S5)
Main author claims¶
- The authors forecast intraday realized volatility using pooled intraday commonality and a market-volatility proxy. (
abstract:S1) - They claim neural networks outperform linear and tree models and remain robust on stocks not included in training. (
abstract:S2,abstract:S3) - They claim proposed forecasts using past intraday RV and time-of-day effects outperform strong daily-RV-only baselines out of sample. (
abstract:S4,abstract:S5)
Data, method, or discussion scope¶
Evidence is confined to abstract claims about model choice and performance, without explicit forecast metric definitions, error distributions, or hyperparameters. (abstract:S1, abstract:S2, abstract:S5)
Main limitations¶
No sensitivity analysis is provided for asset-pool size, intraday partition choices, or non-stationary regime-induced performance degradation. (abstract:S3, abstract:S4, abstract:S5)