How to optimise the code by using condition/loop ?

I have 20 mat files contains two varibles each(let say A,B).

I am calculating linear model between two varibles with fitlm function and I got output like r_square and all

and then I am writing them into TEXT file. I have long code for that

Instead writing each varibles, I want to optimise the code by writing loop/condition

Here is the code

clear;close all
% loading files
load f1_pH_con.mat;load f2_pH_con.mat;load f3_pH_con.mat
load f4_pH_con.mat;load f5_pH_con.mat;load f6_pH_con.mat
load f7_pH_con.mat;load f8_pH_con.mat;load f9_pH_con.mat
load f10_pH_con.mat;load f11_pH_con.mat;load f12_pH_con.mat
load f13_pH_con.mat;load f14_pH_con.mat;load f15_pH_con.mat
load f16_pH_con.mat;load f17_pH_con.mat;
load f18_pH_con.mat;load f19_pH_con.mat;load f20_pH_con.mat
% removing errors(making values NaN if any lesss than zero)
f1_pH(f1_pH<=0)=NaN;f2_pH(f2_pH<=0)=NaN;f3_pH(f3_pH<=0)=NaN;
f4_pH(f4_pH<=0)=NaN;f5_pH(f5_pH<=0)=NaN;f6_pH(f6_pH<=0)=NaN;
f7_pH(f7_pH<=0)=NaN;f8_pH(f8_pH<=0)=NaN;f9_pH(f9_pH<=0)=NaN;
f10_pH(f10_pH<=0)=NaN;f11_pH(f11_pH<=0)=NaN;f12_pH(f12_pH<=0)=NaN;
f13_pH(f13_pH<=0)=NaN;f14_pH(f14_pH<=0)=NaN;f15_pH(f15_pH<=0)=NaN;
f16_pH(f16_pH<=0)=NaN;f17_pH(f17_pH<=0)=NaN;f18_pH(f18_pH<=0)=NaN;
f19_pH(f19_pH<=0)=NaN;f20_pH(f20_pH<=0)=NaN;

f1_con(f1_con<=0)=NaN;f2_con(f2_con<=0)=NaN;f3_con(f3_con<=0)=NaN;
f4_con(f4_con<=0)=NaN;f5_con(f5_con<=0)=NaN;f6_con(f6_con<=0)=NaN;
f7_con(f7_con<=0)=NaN;f8_con(f8_con<=0)=NaN;f9_con(f9_con<=0)=NaN;
f10_con(f10_con<=0)=NaN;f11_con(f11_con<=0)=NaN;f12_con(f12_con<=0)=NaN;
f13_con(f13_con<=0)=NaN;f14_con(f14_con<=0)=NaN;f15_con(f15_con<=0)=NaN;
f16_con(f16_con<=0)=NaN;f17_con(f17_con<=0)=NaN;f18_con(f18_con<=0)=NaN;
f19_con(f19_con<=0)=NaN;f20_con(f20_con<=0)=NaN;
% calculating linear model 
mdl1 = fitlm(f1_pH,f1_con);mdl2 = fitlm(f2_pH,f2_con)
mdl3 = fitlm(f3_pH,f3_con);mdl4 = fitlm(f4_pH,f4_con)
mdl5 = fitlm(f5_pH,f5_con);mdl6 = fitlm(f6_pH,f6_con)
mdl7 = fitlm(f7_pH,f7_con);mdl8 = fitlm(f8_pH,f8_con)
mdl9 = fitlm(f9_pH,f9_con);mdl10 = fitlm(f10_pH,f10_con)
mdl11 = fitlm(f11_pH,f11_con);mdl12 = fitlm(f12_pH,f12_con)
mdl13 = fitlm(f13_pH,f13_con);mdl14 = fitlm(f14_pH,f14_con)
mdl15 = fitlm(f15_pH,f15_con);mdl16 = fitlm(f16_pH,f16_con)
mdl17 = fitlm(f17_pH,f17_con);mdl18 = fitlm(f18_pH,f18_con)
mdl19 = fitlm(f19_pH,f19_con);mdl20 = fitlm(f20_pH,f20_con)

% converting to char for writing into txt file
f1 = evalc('mdl1');f2 = evalc('mdl2');f3 = evalc('mdl3');f4 = evalc('mdl4');
% writing into txt file
fid = fopen(fullfile('regressoion.txt'),'wt');
fprintf(fid,'%s',T);
fclose(fid);

Answer 1

[Rik]

You can start by not using numbered variables.

for n=1:20
    S=load(sprintf('f%d_pH_con.mat',n));
    f_pH{n}=S.(sprintf('f%d_pH',n));%undo the numbered variable mess
    f_con{n}=S.(sprintf('f%d_con',n));%undo the numbered variable mess
    f_pH{n}(f_pH{n}<=0)=NaN;
    f_con{n}(f_con{n}<=0)=NaN;
    mdl{n}=fitlm(f_pH{n},f_con{n});
end

Now you have everything neatly indexed, so you can redesing the way you convert your models to char arrays (not string objects, as your comment claims).

[Jeevan Kumar Bodaballa]

Thank you for your valuble comments and I modified in my question

Can you please tell me how can I covert mdl{n} into char array

[Rik]

You're welcome.

That depends on what exactly you want to store. You can still use evalc, but I'm not a fan, and it is generally not needed.

[Jeevan Kumar Bodaballa]

for example I want store f1_ph,f2_ph,...f20_ph individual varibles likewise others too (f1_con,f2_con,...f20_con)

[Rik]

You should not do that. You see the difference between my code and yours. With my code it is trivial to do the same for 200 files, with your code that is a Herculean task.

What exactly about these models do you want to store in the text file? That is the only part my code doesn't do and yours does.

[Jeevan Kumar Bodaballa]

I understand. Thank you :)