To make vector compatible to matrix.

I have code for disease prediction in which It is equation w = (H'*H+l1*eye(nh,nh))\(H'*Data.y);in which data.y is not compatible for matrix multiplication. As it is vector of labels, either regression or classification labels (+1,-1) so i converted it to matrix by using Data.y=vec2mat(Data.y,297) to make it compatible but again there is same error of inner matrix dimension. What should i do to make it compatible.It is in trainRELM.m . trainRELM is function which is called in main.m I have attched the code and dataset.

Answer 1

[Roger Stafford]

If Data.y is a vector, to make the matrix multiplication H'*Data.y valid, it must be a column vector with the same number of elements as there are columns in H’. That is, size(Data.y) must equal [size(H’,2),1].

[tejasvee]

but H' is of size 100*303 due to size of data set. i have checked the size of H' using size(H'). or you want to tell me that size(H') and size of (Data.y) must be equal.

[Roger Stafford]

No, I didn’t say that. What I have said is that size(Data.y) must equal [size(H’,2),1]. That means that Data.y must have the same number of rows as H’ has columns. Otherwise the expression H'*Data.y will be invalid. Check the site

https://www.mathworks.com/help/matlab/ref/mtimes.html

I also stated that Data.y should have one column because earlier you said that it was a vector. For the purpose of making the above multiplication valid it can actually have any number of columns.

Answer 2

[Walter Roberson]

y is read from columns O P Q R S of the csv file. Those columns are all empty in the supplied csv file so requesting them would temporarily fill them with NaN. If you were using xlsread it would trim away all leading and trailing rows and columns that are all nan which would leave you with emptiness. With csvread it will probably return a 240 x 5 column of nan. However as Roger explained the code can only work when y is a column vector - in this case a column vector with 240 rows .

In the larger scheme even if the number of rows and columns were correct your target information would all be nan and it is not possible to train your classes when you do not have target information.