2
$\begingroup$

What are the best filtering techniques for a highly volatility price series with jumps. What needs to be considered in designing new filters for practical finance which are adaptive and highly reactive to jumps?

Also I have seen a filter called JMA which is proprietary. If I want to create a similar filter, what direction should I take?

$\endgroup$
  • $\begingroup$ You might want to try asking this question on Quantitative Finance. While it's not technically off-topic per se, most people here are signal processing experts/electrical engineers/etc and probably don't have a background in finance. You might have a better shot of an answer there. If you'd still like it to remain here, you should provide more background information so that it makes sense to someone who doesn't know it. $\endgroup$ – Lorem Ipsum Dec 3 '11 at 16:12
  • $\begingroup$ How do I push this to the Quant site. Can I maintain it in both sites. $\endgroup$ – Suminda Sirinath S. Dharmasena Dec 3 '11 at 16:30
  • 1
    $\begingroup$ No, having it on both sites is discouraged and one of them will be deleted. I've messaged a quant mod to ask if they'd be willing to take this question. If you don't want to wait for them to reply, you can simply delete this one and ask on quant yourself and see what the community says. $\endgroup$ – Lorem Ipsum Dec 3 '11 at 16:32
  • $\begingroup$ Lets do the the proper way through the mod. $\endgroup$ – Suminda Sirinath S. Dharmasena Dec 3 '11 at 16:35
  • 1
    $\begingroup$ You're asking for a solution without really giving the question. What do you mean by "adaptive and highly reactive to jumps?" What do you hope for the filter to help you do? Given a highly volatile price series with gaps, what would you like the filter's output to look like? A picture would be helpful. $\endgroup$ – Jason R Dec 5 '11 at 14:44
3
$\begingroup$

In case you're looking for an edge-preserving low pass filter, you could have a look at the bilateral filter.

Implementation is fairly straightforward.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.