Linear Regression in Earth Engine

We learn about the linear regression in earth engine.
This is an attempt to explain the process.
Any comments are welcomed.
---
Codes:
developers.goo...
---
#earthengine
#linearregression

[@alemayehumullatuadashio5350]

Please share with us the code ,it is very interesting tutorial.Much respects!

[@ksabmagar7]

hello sir, all the codes are from the page: developers.google.com/earth-engine/guides/reducers_regression

[@geologistdinesh]

Very interested and ur explanation is highly appreciated thanks sir

[@ksabmagar7]

thankyou

[@Androidkit]

Please make a video on correlation coefficient in Gee

[@ksabmagar7]

sure, i will. thankyou for your kind suggestion.

[@JorgeRodriguez-mp1mt]

Thanks for this video, greetings from Mexico

[@ksabmagar7]

greetings, hope it is helpful.

[@kathahahn20]

Hey there, I have a few questions regarding the proceeding of the linear regression function(s):
- How did you define "system_time:start"? When I use this term I get the notification that it is not defined.
- Do I have to perform all algorithms to perform a Linear Regression? Meaning linearfit and so...?

[@ksabmagar7]

when you previously do other operations, time start properties might be lost, which need to be copied.
no, the functions are specific for linear regression.
Please check the link provided in the description, the video is based on that.
Thankyou for the comment. : )

[@dr.mohanvasave3050]

Very informative video sir.... Please next video upload on Forest fire monitoring

[@ksabmagar7]

Thankyou for the comment.

[@ShaylaTheodore]

Hi！can you tell me how GEE can achieve zero-intercept linear regression (through the origin)?

[@ksabmagar7]

Hello, I don't really know as to be honest.
one way may be to do the standardize the image with the mean and standard deviation - in earth engine, and using it in the linear regression, but this would require number of images in the process. But, not sure if this comment helps.
We can see this discussion too, if you haven't checked:
gis.stackexchange.com/questions/448354/forcing-y-intercept-to-zero-with-google-earth-engine-gee-linear-interpolation
Thankyou

[@ksabmagar7]

Hi, you can share what you tried. I will sure like to see what you did. : )

[@ShaylaTheodore]

@@ksabmagar7 The answer is given in the link you sent me. Set the number of Xnum parameter in the linearRegression function to 1, so that the fitted linear regression model will have no constant term and will definitely pass through the origin.
But I have encountered a new problem, is it possible to perform linear regression with array images? The length of each pixel array is different, so it can't be converted into an image collection for linear regression. It's a headache.😖

[@ksabmagar7]

@@ShaylaTheodore probably due to missing values when masks are applied, may be finding ways to drop any empty values might help in both X and Y variables together. OR may be try printing arrays specifically to see what is missing.

[@ksabmagar7]

@@ShaylaTheodore Hi, so, you solved it all? :)

[@emanuelv5934]

Hi!I have a featureCollection with polygons. Each polygon has a crop yield value associated with it. These polygons are grouped by crop variety. With each polygon I want to calculate a zonal statistic on a spectral index similar to NDVI and perform a linear regression between the yield values ​​and the values ​​obtained from the zonal statistic of the index, but I want to group them by variety and plot the linear regressions for each variety. Could you help me achieve this?

[@ksabmagar7]

you can try computing all the spectral index in earth engine and later export it as csv, which then you can perform in excel/R.
The other way, if you are thinking to do everything in earth engine, might be somewhat learning challenge. It is up to you. You can ask for some help in my page, I will try to help.
Thankyou

[@geologistdinesh]

could you send me one of application developed using this code

[@ksabmagar7]

exactly not ready, but you can use this for finding the trends related study.