MATLAB: Computing rsquared value for data

biological modeldiscrete timersquared

Hi,

I am trying to compute the rsquared for a data set I have; I need two, for each category: juvenile and adult organisms. This is what I have for the RSS and rsquared values:

%Compute RSS y = sum(sum(E1.^2)) + sum(sum(E2.^2))

RSSJ = sum(sum(E1.^2))

RSSA = sum(sum(E2.^2))

%Compute R-squared

residualJ = sqrt(dataJ(2:21,:)) – mean(sqrt(dataJ(2:21,:)));

squareresidualJ = residualJ.^2;

denominator = sum(sum(squareresidualJ));

rsquared = 1 – (RSSJ/denominator)

However, it says that matrix dimensions don't agree for the line containing residualJ. Anyone know what I'm doing wrong?

Best Answer

The mean function will (by default) take the mean of the columns, so the result will be a vector. If you want ‘residualJ’ to be a matrix, use the bsxfun function with the @minus function.

I don’t have your data, so I can’t write specific code, but see if:

residualJ = bsxfun(@minus, sqrt(dataJ(2:21,:)), mean(sqrt(dataJ(2:21,:))));

works. This is obviously UNTESTED CODE.

Related Solutions

MATLAB: Plotting of a prediction line over scatterplot

Sort age before computing y(*); it must be they're not in order so the line follows their sequence in order on the x axis.

() You can, of course, *sort after computing y, saving the order vector optional output and use it to sort the associated y but might as well use the simpler expedient it would seem. If they're already sorted, then the yhat will be those associated with them automagically in the same order (a fairly obvious observation one would presume).

MATLAB: Quantifing goodness of fit between two data sets

If ‘Q’ are your data and ‘Qest’ are the model values, one option is to compute the Coefficient of determination (

) value:

RMSE = sqrt(mean((Q -  Qest).^2));              % Compute RMSE
OLSCF = @(b) sum((Q-Qfcn(b,yez)).^2);
SStot = sum((Q - mean(Q)).^2);
SSres = OLSCF(B);
Rsq   = 1 - (SSres/SStot);                      % Compute R-squared

Best Answer

Related Solutions

MATLAB: Plotting of a prediction line over scatterplot

MATLAB: Quantifing goodness of fit between two data sets

Related Question