How to perform K-means for multidimensional data i.e of size 9*105

Hello,

I am working of Diagnosis of Diseases using RBF. I am using WBCD dataset which can included in MATLAB by "load cancer_dataset". I am trying to use clustering approach for RBFN. But I have no clue to solve multidimensional matrix. Any hints will be thankful.

Answer 1

[Wayne King]

Hi, do you have the Statistics Toolbox? If so see the help for kmeans()

However, kmeans() assumes that rows are observations and columns represent variables or measurements on those observations. It seems from your matrix size that you want to transpose the matrix first assume that you have 105 observations on 9 variables.

It would not be much use to try and cluster 9 observations on 105 variables.

[Manan Nanavati]

Hello Wayne King,

Yes I have Statistics Toolbox.

Its a Breast Cancer Dataset for UCI machine learning repository.

Number of Observation/Patients = 105 & Number of Physical Parameters(variables) = 9.

Considering the syntax of "kmeans()" its kmeans(X,k)

whereas on other side my data are divided also further in to two target instances i.e Cancer 1 and Cancer 2

I want to cluster my data w.r.t. target data i.e Cancer 1 and Cancer 2