Initial commit of LinearRegression class

src/zutil/math/Matrix.java

@@ -4,7 +4,7 @@ package zutil.math;
  * Some basic matrix match functions.
  * Matrix definition: double[y][x].
  */
-public class MatrixMath {
+public class Matrix {
 
     /***********************************************************************
      *                      Scalar
@@ -74,6 +74,8 @@ public class MatrixMath {
      *                      Elemental
      **********************************************************************/
 
+    public static class Elemental {
+
         /**
          * Element addition, each element in matrix1 will be
          * added with the corresponding element in matrix2.
@@ -116,7 +118,7 @@ public class MatrixMath {
          *
          * @return a new matrix with the result
          */
-    public static double[][] elemMultiply(double[][] matrix1, double[][] matrix2){
+        public static double[][] multiply(double[][] matrix1, double[][] matrix2) {
             elementalPreCheck(matrix1, matrix2);
             double[][] result = new double[matrix1.length][matrix1[0].length];
 
@@ -128,17 +130,64 @@ public class MatrixMath {
             return result;
         }
 
+        /**
+         * Element exponential, each element in the vector will raised
+         * to the power of the exp paramter
+         *
+         * @return a new vector with the result
+         */
+        public static double[] pow(double[] vector, double exp) {
+            double[] result = new double[vector.length];
+
+            for (int i = 0; i < vector.length; ++i) {
+                result[i] = Math.pow(vector[i], exp);
+            }
+            return result;
+        }
+
+        /**
+         * Element multiplication, each element in matrix1 will be
+         * multiplied with the corresponding element in matrix2.
+         *
+         * @return a new matrix with the result
+         */
+        public static double[][] pow(double[][] matrix, double exp) {
+            double[][] result = new double[matrix.length][matrix[0].length];
+
+            for (int i = 0; i < matrix.length; ++i) {
+                result[i] = pow(matrix[i], exp);
+            }
+            return result;
+        }
+
         private static void elementalPreCheck(double[][] matrix1, double[][] matrix2) {
             if (matrix1.length != matrix2.length || matrix1[0].length != matrix2[0].length)
                 throw new IllegalArgumentException("Matrices need to be of same dimension: " +
                         "matrix1 " + matrix1.length + "x" + matrix1[0].length + ", " +
                         "matrix2 " + matrix2.length + "x" + matrix2[0].length + ", ");
         }
+    }
 
     /***********************************************************************
      *                      Vector
      **********************************************************************/
 
+    /**
+     * Vector subtraction, every element in the first vector will be subtracted
+     * with the corresponding element in the second vector.
+     *
+     * @return a new vector with subtracted elements
+     */
+    public static double[] subtract(double[] vector1, double[] vector2){
+        vectorPreCheck(vector1, vector2);
+        double[] result = new double[vector1.length];
+
+        for (int i=0; i < result.length; ++i) {
+            result[i] = vector1[i] - vector2[i];
+        }
+        return result;
+    }
+
     /**
      * Matrix Vector multiplication, each element column in the matrix will be
      * multiplied with the corresponding element row in the vector.
@@ -175,6 +224,24 @@ public class MatrixMath {
         return result;
     }
 
+    /**
+     * Sums all values in a vector
+     *
+     * @return the summed value of all elements in the vector
+     */
+    public static double sum(double[] vector) {
+        double sum = 0;
+        for (int i = 0; i < vector.length; i++) {
+            sum += vector[i];
+        }
+        return sum;
+    }
+
+    private static void vectorPreCheck(double[] vector1, double[] vector2) {
+        if (vector1.length != vector2.length)
+            throw new IllegalArgumentException("The two vectors need to have the same length: " +
+                    "vector1 " + vector1.length + "x1, vector2 " + vector2.length + "x1");
+    }
     private static void vectorPreCheck(double[][] matrix, double[] vector) {
         if (matrix[0].length != vector.length)
             throw new IllegalArgumentException("Matrix columns need to have same length as vector length: " +
@@ -196,11 +263,10 @@ public class MatrixMath {
         matrixPreCheck(matrix1, matrix2);
         double[][] result = new double[matrix1.length][matrix2[0].length];
 
-        for (int y=0; y < result.length; ++y) {
+        for (int i=0; i < result.length; ++i) {
-            for (int x=0; x < result[y].length; ++x){
+            for (int k=0; k<matrix1[0].length; ++k) {
-
+                for (int j=0; j < result[i].length; ++j){
-                for (int i=0; i<matrix1[0].length; ++i) {
+                    result[i][j] += matrix1[i][k] * matrix2[k][j];
-                    result[y][x] += matrix1[y][i] * matrix2[i][x];
                 }
             }
         }
@@ -221,6 +287,18 @@ public class MatrixMath {
         return result;
     }
 
+    /**
+     * Sums all values in a matrix
+     *
+     * @return the summed value of all elements
+     */
+    public static double sum(double[][] matrix) {
+        double sum = 0;
+        for (int i = 0; i < matrix.length; i++) {
+            sum += sum(matrix[i]);
+        }
+        return sum;
+    }
 
     private static void matrixPreCheck(double[][] matrix1, double[][] matrix2) {
         if (matrix1[0].length != matrix2.length)

src/zutil/ml/LinearRegression.java

@@ -0,0 +1,38 @@
+package zutil.ml;
+
+import zutil.math.Matrix;
+
+/**
+ * Implementation of a Linear Regression algorithm for "predicting"
+ * numerical values depending on specific input
+ */
+public class LinearRegression {
+
+
+    /**
+     * Method for calculating a hypothesis value fr a specific input value x.
+     * <br><br>
+     * <i>
+     *     h(x) = theta0 * x0 + theta1 * x1 + ... + thetan * xn => transpose(theta) * x
+     * </i>
+     */
+    protected static double[] calculateHypotesis(double[][] x, double[] theta){
+        return Matrix.multiply(x, theta);
+    }
+
+    /**
+     * Linear Regresion cost method.
+     * <br /><br />
+     * <i>
+     *      J(O) = 1 / (2 * m) * Σ { ( h(xi) - yi )^2 }
+     * </i><br>
+     * m = learning data size (rows)
+     * @return a number indicating the error rate
+     */
+    protected static double calculateCost(double[][] x, double[] y, double[] theta){
+        return 1 / (2 * x.length) * Matrix.sum(
+                Matrix.Elemental.pow(
+                        Matrix.subtract(calculateHypotesis(x, theta), y),
+                        2));
+    }
+}

test/zutil/benchmark/AnonymousFunctionBenchmark.java

@@ -0,0 +1,57 @@
+package zutil.benchmark;
+
+import com.carrotsearch.junitbenchmarks.BenchmarkRule;
+import org.junit.Rule;
+import org.junit.Test;
+
+public class AnonymousFunctionBenchmark {
+    public static final int TEST_EXECUTIONS = 500;
+
+    @Rule
+    public BenchmarkRule benchmarkRun = new BenchmarkRule();
+
+
+    private int[] array  = new int[100_000];
+
+
+    @Test
+    public void functionLoop() {
+        for(int k=0; k<TEST_EXECUTIONS; k++) {
+            for (int i = 0; i < array.length; i++) {
+                array[i] = new CalcFunc(){
+                    public int calc(int i){
+                        return i+1;
+                    }
+                }.calc(i);
+            }
+        }
+    }
+
+    @Test
+    public void preFunctionLoop() {
+        CalcFunc func = new CalcFunc(){
+            public int calc(int i){
+                return i+1;
+            }
+        };
+
+        for(int k=0; k<TEST_EXECUTIONS; k++) {
+            for (int i = 0; i < array.length; i++) {
+                array[i] = func.calc(i);
+            }
+        }
+    }
+
+    @Test
+    public void rawLoops(){
+        for(int k=0; k<TEST_EXECUTIONS; k++) {
+            for (int i = 0; i < array.length; i++) {
+                array[i] = i;
+            }
+        }
+    }
+
+    private interface CalcFunc{
+        int calc(int i);
+    }
+}

test/zutil/benchmark/LoopBenchmark.java

@@ -11,30 +11,51 @@ public class LoopBenchmark {
     public BenchmarkRule benchmarkRun = new BenchmarkRule();
 
 
-    private int[] matrix  = new int[100_000];
+    private int[] array1 = new int[100_000];
-    private int[] matrix2 = new int[50_000];
+    private int[] array2 = new int[50_000];
 
 
     @Test
-    public void oneLoop() {
+    public void writeArrayOneLoop() {
         for(int k=0; k<TEST_EXECUTIONS; k++) {
-            for (int i = 0; i < Math.max(matrix.length, matrix.length); i++) {
+            for (int i = 0; i < Math.max(array1.length, array1.length); i++) {
-                if (i < matrix.length)
+                if (i < array1.length)
-                    matrix[i] = i;
+                    array1[i] = i;
-                if (i < matrix2.length)
+                if (i < array2.length)
-                    matrix2[i] = i;
+                    array2[i] = i;
             }
         }
     }
 
     @Test
-    public void twoLoops(){
+    public void writeArraySeparateLoops(){
         for(int k=0; k<TEST_EXECUTIONS; k++) {
-            for (int i = 0; i < matrix.length; i++) {
+            for (int i = 0; i < array1.length; i++) {
-                matrix[i] = i;
+                array1[i] = i;
             }
-            for (int j = 0; j < matrix2.length; j++) {
+            for (int j = 0; j < array2.length; j++) {
-                matrix2[j] = j;
+                array2[j] = j;
+            }
+        }
+    }
+
+
+    @Test
+    public void readArrayLoop() {
+        int sum = 0;
+        for(int k=0; k<TEST_EXECUTIONS; k++) {
+            for (int i = 0; i < array1.length; i++) {
+                sum += array1[i];
+            }
+        }
+    }
+
+    @Test
+    public void readArrayForeach() {
+        int sum = 0;
+        for(int k=0; k<TEST_EXECUTIONS; k++) {
+            for (int i : array1) {
+                sum += array1[i];
             }
         }
     }

test/zutil/math/MatrixTest.java

@@ -7,30 +7,30 @@ import static org.junit.Assert.*;
 /**
  *
  */
-public class MatrixMathTest {
+public class MatrixTest {
 
     @Test
     public void scalarAdd(){
         assertArrayEquals(new double[][]{{4,5},{-2,11}},
-                MatrixMath.add(new double[][]{{2,3},{-4,9}}, 2));
+                Matrix.add(new double[][]{{2,3},{-4,9}}, 2));
     }
 
     @Test
     public void scalarSubtraction(){
         assertArrayEquals(new double[][]{{0,1},{-6,7}},
-                MatrixMath.subtract(new double[][]{{2,3},{-4,9}}, 2));
+                Matrix.subtract(new double[][]{{2,3},{-4,9}}, 2));
     }
 
     @Test
     public void scalarMultiply(){
         assertArrayEquals(new double[][]{{4,6},{-8,18}},
-                MatrixMath.multiply(new double[][]{{2,3},{-4,9}}, 2));
+                Matrix.multiply(new double[][]{{2,3},{-4,9}}, 2));
     }
 
     @Test
     public void scalarDivision(){
         assertArrayEquals(new double[][]{{1,2},{-2,5}},
-                MatrixMath.divide(new double[][]{{2,4},{-4,10}}, 2));
+                Matrix.divide(new double[][]{{2,4},{-4,10}}, 2));
     }
 
 
@@ -38,19 +38,33 @@ public class MatrixMathTest {
     @Test
     public void elementalAdd(){
         assertArrayEquals(new double[][]{{3,5},{-1,13}},
-                MatrixMath.add(new double[][]{{2,3},{-4,9}}, new double[][]{{1,2},{3,4}}));
+                Matrix.Elemental.add(new double[][]{{2,3},{-4,9}}, new double[][]{{1,2},{3,4}}));
     }
 
     @Test
     public void elementalSubtract(){
         assertArrayEquals(new double[][]{{1,1},{-7,5}},
-                MatrixMath.subtract(new double[][]{{2,3},{-4,9}}, new double[][]{{1,2},{3,4}}));
+                Matrix.Elemental.subtract(new double[][]{{2,3},{-4,9}}, new double[][]{{1,2},{3,4}}));
     }
 
     @Test
     public void elementalMultiply(){
         assertArrayEquals(new double[][]{{2,6},{-12,36}},
-                MatrixMath.elemMultiply(new double[][]{{2,3},{-4,9}}, new double[][]{{1,2},{3,4}}));
+                Matrix.Elemental.multiply(new double[][]{{2,3},{-4,9}}, new double[][]{{1,2},{3,4}}));
+    }
+
+    @Test
+    public void elementalVectorPow(){
+        assertArrayEquals(
+                new double[]{4,9,16,81},
+                Matrix.Elemental.pow(new double[]{2,3,-4,9}, 2),
+                0.0);
+    }
+
+    @Test
+    public void elementalMatrixPow(){
+        assertArrayEquals(new double[][]{{4,9},{16,81}},
+                Matrix.Elemental.pow(new double[][]{{2,3},{-4,9}}, 2));
     }
 
 
@@ -59,7 +73,7 @@ public class MatrixMathTest {
     public void vectorMultiply(){
         assertArrayEquals(
                 new double[]{8,14},
-                MatrixMath.multiply(new double[][]{{2,3},{-4,9}}, new double[]{1,2}),
+                Matrix.multiply(new double[][]{{2,3},{-4,9}}, new double[]{1,2}),
                 0.0
         );
     }
@@ -68,18 +82,27 @@ public class MatrixMathTest {
     public void vectorDivision(){
         assertArrayEquals(
                 new double[]{4,1},
-                MatrixMath.divide(new double[][]{{2,4},{-4,10}}, new double[]{1,2}),
+                Matrix.divide(new double[][]{{2,4},{-4,10}}, new double[]{1,2}),
                 0.0
         );
     }
 
+    @Test
+    public void vectorSum(){
+        assertEquals(
+                20.0,
+                Matrix.sum(new double[]{1,2,0,3,5,9}),
+                0.02
+        );
+    }
+
 
 
     @Test
     public void matrixMultiply(){
         assertArrayEquals(
                 new double[][]{{486,410.4,691.6},{314,341.6,416.4},{343.5,353.4,463.6},{173,285.2,190.8}},
-                MatrixMath.multiply(
+                Matrix.multiply(
                         new double[][]{{1,2104},{1,1416},{1,1534},{1,852}},
                         new double[][]{{-40,200,-150},{0.25,0.1,0.4}})
         );
@@ -89,21 +112,30 @@ public class MatrixMathTest {
     public void matrixTranspose(){
         assertArrayEquals(
                 new double[][]{{1,3},{2,5},{0,9}},
-                MatrixMath.transpose(
+                Matrix.transpose(
                         new double[][]{{1,2,0},{3,5,9}})
         );
     }
 
+    @Test
+    public void matrixSum(){
+        assertEquals(
+                20.0,
+                Matrix.sum(new double[][]{{1,2,0},{3,5,9}}),
+                0.02
+        );
+    }
+
 
 
     @Test
     public void identity(){
         assertArrayEquals(
                 new double[][]{{1}},
-                MatrixMath.identity(1));
+                Matrix.identity(1));
 
         assertArrayEquals(
                 new double[][]{{1,0,0,0},{0,1,0,0},{0,0,1,0},{0,0,0,1}},
-                MatrixMath.identity(4));
+                Matrix.identity(4));
     }
 }