kmeans on 3d data

Hi,

I have 3d data with 24*2*705 matrix. i would like to do k-means clustering on this data.

I am not able to run kmeans command as it is showing error as "Transpose data into column orientation".

I tried with permute but not working.

Can anybody help on how to handle 3d data for kmeans clustering.

Regards

Vaishali

Answer 1

[Image Analyst]

So, did you transpose it like it suggested? It wants 3 columns and N rows, not 3 rows and N columns like you have. Attach your data if you need any more help.

[Vaishali Nagpure]

Hi,

Have you tried on my data?

Regards

Vaishali

[Image Analyst]

I just read it in but I don't know how to interpret this data. What do each of the 3 dimensions represent, (x,y,z)? What does the value at that location, at data3d(index1, index2, index3), represent?

[Vaishali Nagpure]

Hi,

24*2 represents mean(first column) and coefficient of variance (second column) of hourly data. It is 24 hrs data. And third dimension is data for 705 frequencies. So for 705 frequencies, 24 hours means (first columns) and coefficient of variance (2nd column).

Regards

Vaishali

[Image Analyst]

But the data is not in a 2-Dimensional N rows-by-3 columns matrix. It's a 3-D array. So there are not 3 columns like you said. Please try to explain again.

[Vaishali Nagpure]

There are no 3 columns. Data is 24*2 matrix. So 24 rows and two columns. rows are hours. So 24 hours. First column is mean. Second column is coefficient of variance(cv). So 24*2 data represent data as: 24 hours mean and cv.

This data is for 705 frequencies. So 24*2*705 data. For each frequency 24*2 data.

Regards

Vaishali

Answer 2

[Vaishali Nagpure]

Please find my data.

Regards

Vaishali