[R-SIG-Finance] Online Hayashi-Yoshida estimator

[Robert A'gata]

Hi Brian - To the best of my knowledge, I don't know any literature that
describes an on-line estimation method for the HY covariance estimator. I
am raising this question on this forum to brainstorm and get some ideas.

Intuitively since the estimator is only caring about overlapping returns
between the two asynchronous tick streams, doing this online might not seem
to be a daunting task. I attempted a couple of times by trying to keep
tracking of last overlapping period and discard all the ticks that are
completely "overlapped or enveloped" in the past. But for some reason, it
didn't go so well. So I would like to pick some ideas if anybody has any.

Thank you.

Robert

On Tue, Mar 17, 2015 at 4:29 AM, Brian G. Peterson <brian at braverock.com>
wrote:

> Do you have a reference for the on-line version?  It generally helps
> others to help you if they can look at the paper(s) which describe the
> technique you wish to replicate.
>
> Regards,
>
> Brian
>
> On 03/16/2015 11:39 PM, Robert A'gata wrote:
>
>> Thank you Hasan. I looked into the c code of Yuima. It is an off-line
>> version, i.e. it passes in entire arrays of returns on both assets for
>> which you want to compute covariance. Rather, I'm looking for an idea to
>> implement the estimator in an on-line fashion.
>>
>> On Tue, Mar 17, 2015 at 12:04 AM, Hasan Diwan <hasan.diwan at gmail.com>
>> wrote:
>>
>>  There is a yuima package that claims to have the Hayashi-Yoshida
>>> covariance
>>> estimator since 2009. The link is
>>> http://r-forge.r-project.org/projects/yuima/ and I hope it helps. -- H
>>>
>>> On 16 March 2015 at 20:19, Robert A'gata <rhelpacc at gmail.com> wrote:
>>>
>>>  Hi,
>>>>
>>>> Does anyone know of an on-line implementation of Hayashi Yoshida
>>>>
>>> covariance
>>>
>>>> estimator? Let's say that we have to time series ticked asynchronously
>>>>
>>> and
>>>
>>>> we would like to update the realized covariance once either ticks so
>>>> that
>>>> we do not have to keep in memory all the tick history. Any idea please?
>>>> Thank you.
>>>>
>>>> Best,
>>>>
>>>> Robert
>>>>
>>>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
> should go.
>

	[[alternative HTML version deleted]]