Predict exact value by using Pre trained GPR Model(gaussian process regression)

i have 4 input data(a,b,c,d) and 1 output data.

%% MGPR(Multivariate Gaussian process regression) by using metamodel %%

clc; clear all; close all; warning off;

tbl = readtable('D-optimal.txt','Filetype','text',...

'ReadVariableNames',false);

tbl.Properties.VariableNames = {'Hole_horizon','Hole_vertical','Height','Depth','Deflection'};

tbl(1:20,:) %Data

gprMdl1 = fitrgp(tbl,'Deflection','KernelFunction','ardsquaredexponential',...

'FitMethod','sr','PredictMethod','fic','Standardize',1)

gprMdl2 = fitrgp(tbl1,'Deflection','KernelFunction','ardsquaredexponential',...

'FitMethod','sr','PredictMethod','fic','Standardize',1)

i wanna predict the exact value of deflection(Output).

is there any way to predict the output value about random input value?

Answer 1

[Ayush Anand]

Hi JaeWook,

I understand you have trained a Gaussian process model on some data and want to know how to use it to predict the values for certain input data you have. You can do this using the "predict" function, with the following syntax:

ypred=predict(gprMdl1,xtest) %where xtest is your test input data

You can refer to the following example which demonstrates how to train a Gaussian process regression model and then use it to predict values at certain inputs:

https://in.mathworks.com/help/stats/predict-battery-soc-using-machine-learning.html

I hope this helps!