RicardoSantos

2021年4月4日晚上9點27分

Function - Kernel Density Estimation (KDE)

Bitcoin / U.S. dollarBitstamp

描述

⋅ 2021年4月4日晚上9點27分

"In statistics, kernel density estimation (KDE) is a non-parametric way to estimate the probability density function of a random variable."
from wikipedia.com

KDE function with optional kernel:

Uniform
Triangle
Epanechnikov
Quartic
Triweight
Gaussian
Cosinus

Republishing due to change of function.
deprecated script: tradingview.com/script/vWOxbhhj-KDE-Gaussian/

發布通知

⋅ 2021年4月4日晚上9點45分

added quartic and triweight kernels.

發布通知

⋅ 2021年4月5日上午10點58分

added placeholder for kernels(logistic, sigmoid, silverman)
added kernel calculations for kernel(uniform, triangular, cosine)

發布通知

⋅ 2021年4月5日下午2點24分

added calculations for kernels(logistic, sigmoid and silverman(Not working atm)

發布通知

⋅ 2021年4月7日上午10點06分

removed silverman kernel, added highest value index line/label, nearest to 0 index as a dotted gray line.

發布通知

⋅ 2021年4月21日上午9點32分

added extra stats/visuals to drawing function.

評論

kakola

⋅ 2021年4月6日早上12點09分

That bandwidth is phire. 🔥
*phi*re

Yelian

⋅ 2023年1月13日上午10點45分

Is it just me or does the script provide the same output regardless of the chart symbol?

Yelian

⋅ 2023年1月8日下午3點19分

Can someone please provide some tips on how one may use this awesome script?

kakola

⋅ 2021年10月30日下午2點19分

Can be updated to use pinescript's new "switch" function.

RicardoSantos

⋅ 2021年10月30日下午3點12分

@kakola, this is now deprecated, V5 library version can be found here:
tradingview.com/script/oTcbPUuo-MathStatisticsKernelDensityEstimation/

THE_REAL_BCT

⋅ 2021年7月9日下午1點22分

Nice little script... With one point I don't quite see how it works. Supposedly arrays can store up to 100K items. For whatever reason, I can't get more than 7 or 8 floats pushed into a float array or Pine blows up. So... That's not a whole lot of obs to infer a pdf

As a matter of fact I see you have 7 obs preloaded.

How do you load up obs beyond 7 ?

I have an idea for how to circumvent that limitations (using int array - altho I think they blow up at 50 or so) but want to see if you have something better? or maybe it's just a user/noob error and none of you guys see this problem?

RicardoSantos

⋅ 2021年7月9日下午1點29分

@hypercrush1, yes the high complexity of the script limits its utility alot, unfortunely..

THE_REAL_BCT

⋅ 2021年7月9日下午1點37分

@RicardoSantos, gotit. thanks. Love the code you produce!

BTW I just tried a script that does nothing else than push 'close' into a float array... then do 2 things (sequentially)
1/ ping array.size and display it, and it returns 1300+ or so - scrip executes fine
2/ get item before last pushed on the array/stack - Pine pukes. Error message "array size = 1" (so it looks like pine lied in 1/ and all along!)

Is that a Pine bug? or feature?

THE_REAL_BCT

⋅ 2021年7月9日下午2點11分

@THE_REAL_BCT, Ignore the last comment - user error ;)

KB0006

⋅ 2021年5月12日早上1點46分

Thanks for sharing this, I've been trying to find an indicator that uses kernel method and this seems to be the only one! Though I don't quite understand what those numbers suggest. Can you explain it to a dummy like myself? lol