Activity 20: Neural Networks

For this activity, we classify the same set of objects as with activities 18 and 19 but using neural networks instead. A neural network will have three parts, the input layer, the hidden layer and the output layer. For classification purposes, the input layer contains the features and the output would contain the classes. As with activity 19 I used piatos and pillows as my 2 classes. The output may either be 0 (piatos) or 1 (pillows) depending on which class a sample belongs.
The code below follows the one written by Jeric. I used 4 input layer neurons, 8 hidden layer neurons and 1 output layer neuron. The following is the code I used, the training and test sets are the ones in activity 19.
-----------------------------------------------------------------------------------
chdir("C:\Users\RAFAEL JACULBIA\Documents\subjects\186\activity20");
N = [4,8,1];

train = fscanfMat("train.txt")';// Training Set

t = [0 0 0 0 1 1 1 1];
lp = [2.5,0];
W = ann_FF_init(N);

T = 500; //Training cyles
W = ann_FF_Std_online(train,t,N,W,lp,T);

test = fscanfMat("test2.txt")';

class = ann_FF_run(test,N,W)
round(class)
--------------------------------------------------------------------------

The following are the outputs of the program :
0.0723331 0.0343279 0.0358914 0.0102084 0.9948734 0.9865080 0.9895211 0.9913624

After rounding off, we see that the above outputs are 100% accurate.

I grade myself 10/10 for again getting a perfect classification. Thanks to jeric for the code and to ed for his help.