Edge effect on the way to the GRAAL - page 11
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I have a regression model, but have never thought to look into estimating it,
The wavelet is messing up too.
This forum attracted me to the 'edge effect' in wavelets. It's a known problem. We fit a model on a sample and then extrapolate it "over the edge of the sample" and it doesn't work there. If we solve the edge effect in wavelets, will we be able to solve the problem of extrapolating the model out of sample?
the wavelet is messed up too.
This forum attracted me to the "edge effect" in wavelets. It's a known problem. We fit a model on a sample and then extrapolate it "over the edge of the sample" and it doesn't work there. If we solve the edge effect in wavelets, will we be able to solve the problem of extrapolating the model out of sample?
How do you solve it?
Once again: this is the fundamental problem of the wavelet. That is, if you "solve" it, it won't be a wavelet at all.
How do you solve it?
Once again: this is a fundamental problem of the wavelet. That is, if you "solve" it, it won't be a wavelet at all.
The point is that it's initially a bit absurd to analyse only onlk
What else? Rumours, news?
The point is that it's initially a bit absurd to analyse only onlk
What else? Rumours, news?
Why not? News has a lot to do with it. Trends often break, for example.
What's more - 70% of the "important news" is just numbers. Changes in macroeconomic indicators.
And why not? The news does have an impact. Trends often break, for example.
What's more - 70% of "important news" is just numbers. Changes in macroeconomic indicators.
You're kidding.
I'm kind of at a loss as to what to say.
But actually, no. I'm not joking.
If you pull out, for example, the calendar for last year, break it down into machine-readable form, manually throw out all sorts of holidays and stuff, and overlay it on the chart, you'll see a lot of interesting things.
I'm kind of at a loss as to what to say.
But actually, no. No kidding.
For example, if you pull out the calendar for last year, break it down into machine-readable form, manually throw out all kinds of holidays and so on, and overlay it on the chart, you'll see a lot of interesting things.
There is the science of measuring economic data. And there is shamanism with shamanism.
And then there are the nonsensical phrases. Meaningless in their purported meaningfulness.
What were you trying to say?