I suggest

Handbook of Financial Time Series

Editors: Andersen, T.G., Davis, R.A., Kreiss, J.-P., Mikosch, Th.V. (Eds.)

Thanks @Babak Jamshidi

Financial time series mostly show volatility, which is a tendency to fluctuate erratically. Such a tendency is associated with non-constancy of variance across the time axis. There is emergence of variance models of the arch-garch family to model these time series.

@Ette Etuk Thanks for your Opinion.

I suggest to look at the survey paper, which I am a co-author:

https://arxiv.org/pdf/2012.03854.pdf

You will find there a brief description of methods, models and references to the literature.

@Piotr Fiszeder thanks for sharing this useful material here. I shall highly appreciate if you can share anything on recent development of volatility analysis and Forecasting.

I think usually, we use a model to forecast the regular trend, and the difference between real price and the forecast one, or, the residual, shows the volatility.

@Qian Wang, Thank you.

Dear Fakir Tajul Islam,

Rough volatility models, in which the volatility is modelled by a fractional Brownian motion (that is a Brownian motion with a particular dependence among increments, parametrized by the Hurst exponent) make it possible to forecast volatilities.

• The most famous paper introducing the concept on stocks (a section is devoted to forecast): https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.745.6660&rep=rep1&type=pdf

• An extension of this approach also to exchange rates:

  Preprint Long vs Short Time Scales: the Rough Dilemma and Beyond

Great discussion