Wrong with matrix dimensions

I'm only a beginner when it comes to MatLab and would therefor appreciate some help.

This is my code:

clc

clear

close all

load geiger.txt;

load temperatur.txt; %hämtar filerna med relevant data

i = geiger(:,4);

j = i*4500;

t1 = geiger(:,1);

x = [1:120]; %omvandling till rätt vektor

tidW = (x-1)*10;%omvandling till rätt enhet

t2 = temperatur(:,1);

tidT = (t2-1)*10;

T = temperatur(:,4)+273;

acc(1) = j(1);

for h = 2:length(j) %lägger ihop pulserna från geigermätaren

    acc(h)= acc(h-1)+j(h);

end

p1 = polyfit(tidW,acc,3); %anpassar polynom till W(t)

pv1 = polyval(p1,tidW); %gör om värdena med polynomet som grund

pd1 = polyder(pv1); %deriverar polynomet

p = polyval(pd1, tidW)

p2 = polyfit(t2,T,2); %anpassar polynomet T(t)

pv2 = polyval(p2,t2);

pd2 = polyder(p2);

p2 = polyval(pd2,t2);

PH = 10*4190*p2;

e = PH./p;

plot(tidW,acc,'r')

hold on

plot(tidW, pv1,'b')

figure(2)

plot(t2, T,'r')

hold on

plot(t2,pv2,'b')

figure(3)

plot(t2,e)

And MatLab tells me that there's wrong with the matrix dimensions (matrix dimensions must agree)

This is due on tuesday and I'm getting a bit stressed out! All sort of help is highly appreciated!

Answer 1

[Matt Tearle]

I assume the error is at the bolded line ( e = PH./p ). It looks like this is due to a row vs column issue (as Walter mentions).

PH comes from p2, which comes from t2, which appears to be a column vector ( t2 = temperatur(:,1); ). So PH should be a column.

p however comes from tidW, which comes from x, which is a row vector ( x = [1:120]; ).

So you're trying to do an element-wise divide of a column by a row. No can do. Try simply transposing x to a column: x = (1:120)';

[Lisa]

Thanks, it helped a lot. Unfortunately new errors occured instead.

It says now that the vectors has to be the same size. I interpret this as there has to be the same amount of values in the two vectors, which there is.

[Matt Tearle]

Maybe, but not necessarily. Size in MATLAB means both rows and columns, so a row vector and column vector are different sizes, even if they have the same number of elements.

Can you please copy and post the error message you're getting, so we can see exactly which variables and which line of code it's referring to.

Also, please post the output from the command >> whos This is a useful diagnostic, if you don't know of it. It shows all your variables and their sizes.

Size matching issues are probably about the second most common error messages in MATLAB, so it's good to know how to track them down. Interactively, you can look at the workspace. At the command line, use the whos command.

[Lisa]

??? Error using ==> polyfit at 48

X and Y vectors must be the same size.

Error in ==> varmepump2 at 24

p1 = polyfit(tidW,acc,3); %anpassar polynom till W(t)

Name Size Bytes Class Attributes

T 120x1 960 double

acc 1x120 960 double

geiger 120x4 3840 double

h 1x1 8 double

i 120x1 960 double

j 120x1 960 double

t1 120x1 960 double

t2 120x1 960 double

temperatur 120x5 4800 double

tidT 120x1 960 double

tidW 120x1 960 double

[Matt Tearle]

Yep, same issue: acc is a row, tidW is a column. See my other answer for a better way to calculate acc that also keeps it the same size as j (or just transpose it).

[Lisa]

Okay. Thanks! It is now kind of working. Might be some trouble with a graph, but I'll check my calculations because I think the problem's there!

Once again, thank you!

[Matt Tearle]

You're very welcome. Can you accept and/or vote for answers that helped, so others know this has been solved (and what the solution was). Thanks. Open a new question if you keep having problems with the plot.

Answer 2

[Walter Roberson]

Which line is it complaining about, and what are the sizes of the variables involved?

I notice that most of your variables are column vectors. However, because your acc array does not have a preallocated size, it will default to being a row vector. There are certainly cases where the difference between row vector and column vector could cause problems.

Answer 3

[Matt Tearle]

Aside: you can replace

acc(1) = j(1);
for h = 2:length(j) %lägger ihop pulserna från geigermätaren
    acc(h)= acc(h-1)+j(h);  
end

with

acc = cumsum(j);

Oh yeah baby!