working state

.gitignore

@@ -0,0 +1 @@
+.octave-workspace

activation_tanh.m

@@ -0,0 +1,4 @@
+function [f,fder] = activation_tanh(b)
+   f = @(x) tanh(b*x);
+   fder = @(x) b*(1-f(x).^2);
+end

multilayer_perceptron.m

@@ -0,0 +1,43 @@
+2;
+function output = get_output(entry,weights,activation_func)
+    output = activation_func(weights * entry);
+end
+
+function [weights,output] = multilayer_perceptron_learn(entries, expected_output, neurons_per_layer, activation_func, activation_der,
+                                    learning_factor=.5, max_iterations=1000, tolerance=1e-5)
+    n = length(entries(:, 1));
+    m = 0;
+    for i = 2:length(neurons_per_layer)
+        m = m + 1;
+        weights{m} = rand(neurons_per_layer(i), neurons_per_layer(i-1)+1) .- 0.5;
+        layer_entry{m} = [-1, zeros(1, neurons_per_layer(i-1))];
+    end
+    M = m
+    for iteration = 1:max_iterations
+        for index = randperm(2^n); 
+                layer_entry{1}(2:end) = entries(:, index);
+            for m = 2:M
+                layer_entry{m}(2:end) = activation_func(weights{m-1} * layer_entry{m-1}');
+            end
+            layer_entry;
+            output(index) = activation_func(weights{M} * layer_entry{M}');
+            output(index);
+            d{M} = activation_der(weights{M}*layer_entry{M}')*(expected_output(index) - output(index));
+            for i = M-1:-1:1
+                d{i} = activation_der(weights{i}*layer_entry{i}')*(weights{i+1}(:,2:end)*d{i+1});
+            end
+            d;
+            weights;
+            for i = 1:M
+                delta_w = learning_factor * layer_entry{i}'*d{i};
+                weights{i} = weights{i} + delta_w';
+            end
+        end
+        sum(abs(expected_output - output))
+        if sum(abs(expected_output - output)) <= tolerance
+            return
+        end
+    end
+
+end
+