How to select thresholds on a XY plot and fit a line btw thresholds

Hi there, I have a matrix (test & test1) of vectors that I split into a matrix of X (ANG) values and a matrix of Y (MOM) values. I am plotting XY in a scatter plot. I then want to identify when the slope of the XY plot is changing. Those points will be the thresholds I need.

I want the code to be able to identify thresholds in both cases. This is what I have done:

%%Identify index for 5% and 95% MOM using threshold

Mom5up_indx=zeros(1,length(MOM));
Mom95up_indx=zeros(1,length(MOM));
Mom5down_indx=zeros(1,length(MOM));
Mom95down_indx=zeros(1,length(MOM));

for i=1:length(Mom5)

    Mom5up_indx(1,i)= find(MOM(:,i)>=Mom5(1,i),1,'first');
    Mom95up_indx(1,i)= find(MOM(:,i)>=Mom95(1,i),1,'first');
    Mom5down_indx(1,i)= find(MOM(:,i)>=Mom5(1,i),1,'last');
    Mom95down_indx(1,i)= find(MOM(:,i)>=Mom95(1,i),1,'last');

end

%%Identify Linear fit and Slope value for the three segments of the ANG-MOM signal

%Define thresholds

THR1= Mom5up_indx;    % First crossing 5% of Max MOM
THR2= ??;             % First change in Slope after 5% of Max MoM
THR3= ??;             % Second change in Slope after 5% of Max MoM
THR4= Mom95up_indx;   % First crossing 95% of Max MOM
THR5= Mom95down_indx; % Last before crossing 95% of Max MOM
THR6= Mom5down_indx;  % Last before crossing 5% of Max MOM

% Preallocation of variables

Slope1=zeros(1,length(ANG));
Slope2=zeros(1,length(ANG));
Slope3=zeros(1,length(ANG));

for i=1:length(ANG)

   f= fit(ANG([THR1(i):THR2(i)],i),MOM([THR1(i):THR2(i)],i),'poly1');
   Slope1(1,i) = f.p1(1);

   f1= fit(ANG([THR3(i):THR4(i)],i),MOM([THR3(i):THR4(i)],i),'poly1');
   Slope2(1,i) = f1.p1(1);

   f2= fit(ANG([THR5(i):THR6(i)],i),MOM([THR5(i):THR6(i)],i),'poly1');
   Slope3(1,i) = f2.p1(1);

end

Hope you have all the info you need to help me. Many thanks

Ale

[dpb]

Instead of (or in addition too) attaching .fig files, how about saving as .jpg and inserting images so folks can see w/o downloading?

[Alessandro Garofolini]

Thanks dpb! good suggestion ;)

Ale

Answer 1

[Nawaf Almutairi]

hud