add learn multiple out

Author: asm-jaime

MLTests/GradientOneInMulOutTests.cs

@@ -0,0 +1,86 @@
+using FluentAssertions;
+using NUnit.Framework;
+
+namespace MLTests
+{
+    public class GradientOneInMulOutTests
+    {
+        private double[] ElementMul(double[] vecA, double element)
+        {
+            var result = new double[vecA.Length];
+            for (int i = 0; i < vecA.Length; i++)
+            {
+                result[i] = vecA[i] * element;
+            }
+            return result;
+        }
+
+        private double[] Mul(double[] vecA, double[] vecB)
+        {
+            var result = new double[vecA.Length];
+            for (int i = 0; i < vecA.Length; i++)
+            {
+                result[i] = vecA[i] * vecB[i];
+            }
+            return result;
+        }
+
+        private double[] Sub(double[] vecA, double[] vecB)
+        {
+            var result = new double[vecA.Length];
+            for (int i = 0; i < vecA.Length; i++)
+            {
+                result[i] = vecA[i] - vecB[i];
+            }
+            return result;
+
+        }
+
+        private double[] SubA(double[] vecA, double[] vecB, double alpha)
+        {
+            var result = new double[vecA.Length];
+            for (int i = 0; i < vecA.Length; i++)
+            {
+                result[i] = vecA[i] - vecB[i]*alpha;
+            }
+            return result;
+
+        }
+
+        private double[] GradientLearnByOneInputToMulOut(double[] weights, double input, double[] predictionGoal, double alpha, int iterations)
+        {
+            var result = weights.ToArray();
+            for(int i = 0; i < iterations; i++)
+            {
+                var predictions = ElementMul(result, input);
+                var deltas = Sub(predictions, predictionGoal);
+                var deltasWeights = ElementMul(deltas, input);
+                result = SubA(result, deltasWeights, alpha);
+            }
+            return result;
+        }
+
+        [Test]
+        public void ShouldGradientLearnMulOut()
+        {
+            var weights = new double[] { 0.3, 0.2, 0.9 };
+
+            var wlrec = new double[] { 0.65, 1.0, 1.0, 0.9 };
+
+            var hurt  = new double[] {0.1, 0.0, 0.0, 0.1};
+            var win = new double[] { 1, 1, 0, 1 };
+            var sad = new double[] { 0.1, 0.0, 0.1, 0.2 };
+
+            var alpha = 0.2;
+            var iterations = 3000;
+
+            var input = wlrec[0];
+            var predictionGoal = new double[] { hurt[0], win[0], sad[0] };
+            var weightsLearned = GradientLearnByOneInputToMulOut(weights, input, predictionGoal, alpha, iterations);
+            var result = ElementMul(weightsLearned, input);
+            Assert.That(predictionGoal[0], Is.EqualTo(Math.Round(result[0], 2)));
+            Assert.That(predictionGoal[1], Is.EqualTo(Math.Round(result[1], 2)));
+            Assert.That(predictionGoal[2], Is.EqualTo(Math.Round(result[2], 2)));
+        }
+    }
+}