How to randomly divide two time-series of input values and target values into training, validation and test sets?

I am new to Matlab.

I have my time-series of input values x (220*1) and target values y (220*1) that are resembled in two seperate Workplaces. I want to randomly divide (x,y) into trainning, validation and test set in proportion of (0.8, 0.1, 0.1) to find the best degree of my Polynomial Curve Fitting. However, I couldn't follow the instruction by https://www.mathworks.com/help/deeplearning/ref/dividerand.html since my data are in seperate sheet, and I don't know how to write the codes achieving such purpose with each y values corresponds to the x values that are in the same row.

Could you please show me the codes of doing that?

Thank you so much!

Answer 1

[KSSV]

x = 1:220 ; 
y = rand(size(x)) ; 
Q = [x' y'] ; 

trainRatio = 70/100 ; 
valRatio = 15/100 ; 
testRatio = 15/100 ;

[trainInd,valInd,testInd] = dividerand(length(x),trainRatio,valRatio,testRatio) ; 

training = Q(trainInd,:) ;
validation = Q(valInd,:) ; 
testing = Q(testInd,:) ; 

[Robin Li]

Thank you! It works! :)

[MatLabMcLovinPotato]

Curious please - wondering if this could be changed to be:

• not random
• for 3 (not one) input
• a timeseries (can / does data data some into training?)