How to create target matrix?

I have extracted 21 features of 12 samples. I need to create a target matrix. Plz help.

[Jan]

I recommend to explain any details if you want others to understand, what you are doing.

[KSSV]

Your target depends on what is your problem...what are features..what algorithm you are going to use. If the problem is of clustering, you may not have target itself. With out this information tough to get help. Explain more.

[Walter Roberson]

You should start with a target vector of length 12, in which the K'th entry tells you which class sample #K belongs to.

Depending on the routine you use, you might need to use ind2vec()

[Greg Heath]

Considering the lack of details in your question, I'm surprised you received any responses. It is not even clear whether this is regression or classification.

The more info you choose to reveal, the more likely the responses will help.

Greg

[Aya Ahmed]

if It is classification. what we need to create a target matrix ?

[Walter Roberson]

You need target information in order to do the training.

In machine learning, there are two kinds of algorithms:

1. algorithms which start with some information about what the "right" answer is for some cases, and use it to figure out useful characteristics to separate the various cases correctly; or
2. algorithms which are not given any information about what samples belong to what, and have to somehow figure out what the groupings are anyhow.

The second kind of algorithm is potentially useful for organizing information when no organization is known ahead of time, or for discovering organizing factors that had been previously overlooked. However, it is much more difficult to get useful answers out of in a reasonable time.

Answer 1

[Greg Heath]

Considering the lack of details in your question, I'm surprised you received any responses. It is not even clear whether this is regression or classification.

The more info you choose to reveal, the more likely the responses will be relevant.

If you have N I-dimensional "I"nputs that are paired with N O-dimensional "O"utput targets for designing a net with O-dimensional outputs

 [ I N ] = size(input)

 [ O N ] = size(target) % = size(correspondingoutput)

Hope this helps.

Thank you for formally accepting my answer

Greg