The parameter 'Y' must be a real-valued vector with length same as the columns of 'X'.

I have a 447x10 matrix, which I subdivided into a 447x1 vector of an Y-response variable and a 447-9 vector of a X-predictor variables. I get the above error when I try:

[B,FitInfo] = lasso(Y,X);

Using n = length(Y) yields the same length for Y and X, while

tf = isreal(Y) and 
tf = isreal(X)

both return 1, which means both Y and X are real-valued vectors with the same length. Anything I overlooked? Thanks

[dpb]

Show, don't just tell :)

"which means both Y and X are real-valued vectors..."

Not precisely, no. Means neither contain an imaginary part but says nothing about shape.

You keep using "vector", but X needs to be an array.

What does

whos Y X

return?

[Walter Roberson]

lasso(X,Y) -- you have the wrong order

Answer 1

[Peter_SG]

Thank you very much for all your replies - I had indeed had the wrong order. Thanks and apologies!