## How to interpret an answer given by the Neural Network

on 11 Dec 2012

### Greg Heath (view profile)

Hello people :)

Before I begin, I want to say that I am a beginner in the Neural Network field so, if the question seems stupid... please give me a correct answer :)

Yesterday I began to experiment with the Neural Network tools, more precisely with nnstart and nftool. I started with the famous house pricing example. I did everything the tutorial told me to do and I got an answer consisting in a matrix with one row and 506 columns, stored in the output variable.

The first three elements are as follows:

1. 24.6073
2. 22.6764
3. 34.6302

And so on. So far so good. But here are my questions:

1. How to interpret those results? What should I understand from them? That the median price for neighborhood 1 should be or will be 24.6073? Or how?
2. Let's say that I have a 507th set of data. How can I find the answer for that single set of data, based on the data that the NN already has, without entering it into the input variables, re-training the network, etc.?

### Greg Heath (view profile)

on 12 Dec 2012
Edited by Greg Heath

### Greg Heath (view profile)

on 12 Dec 2012

Q1. How to interpret those results? What should I understand from them? That the median price for neighborhood 1 should be or will be 24.6073?

A1. For a house with those 13 input parameters, the predicted median house price is \$24,607.30

Q2.Let's say that I have a 507th set of data. How can I find the answer for that single set of data, based on the data that the NN already has, without entering it into the input variables, re-training the network, etc.?

A2. The purpose of training the model is to estimate a median price given the 13 input parameters. Since the model is trained, to estimate the price of other houses not in the original set of 506, just use

priceestimate = net(newinput)

Otherwise design another model.

Hope this helps.

Thank you for formally accepting this answer.

Greg

### Bogdan (view profile)

on 12 Dec 2012

Thank you very much :)

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi