Changes between Version 32 and Version 33 of Documentation/Reference/GPDL

Timestamp:

11/02/13 09:51:25 (10 years ago)

Author:

gkronber

Comment:

--

Documentation/Reference/GPDL

@@ -87 +87 @@
   double[,] x;
   double[] y;
+  int[] rows;
   string[] variableNames;
+  double[] randomConsts;
+  
   Dictionary<string,int> nameToCol;
   
@@ -102 +105 @@
 
   double RSquared(IEnumerable<double> xs, IEnumerable<double> ys) {
-     HeuristicLab.Problems.DataAnalysis.OnlineCalculatorError error;
-     var r2 = HeuristicLab.Problems.DataAnalysis.OnlinePearsonsRSquaredCalculator.Calculate(xs, ys, out error);
-     if(error == HeuristicLab.Problems.DataAnalysis.OnlineCalculatorError.None) return r2;
-     else return 0.0;
-  }
+    // calculate Pearson's correlation in one pass over xs and ys
+    double sumx = 0.0;
+    double sumy = 0.0;
+    double sumxSq = 0.0;
+    double sumySq = 0.0;
+    double sumxy = 0.0;
+    int n = 0;
+    var xEnum = xs.GetEnumerator();
+    var yEnum = ys.GetEnumerator();
+    while(xEnum.MoveNext() & yEnum.MoveNext()) {
+      sumx += xEnum.Current;
+      sumy += yEnum.Current;
+      sumxSq += xEnum.Current * xEnum.Current;
+      sumySq += yEnum.Current * yEnum.Current;
+      sumxy += xEnum.Current * yEnum.Current;
+      n++;
+    }
+    System.Diagnostics.Debug.Assert(!(xEnum.MoveNext() | yEnum.MoveNext()));
+
+    double num;
+    double den;
+    double r = 0.0;
+    num = sumxy - ( ( sumx * sumy ) / n );
+    den = Math.Sqrt( ( sumxSq - ( sumx*sumx ) / n ) *
+                     ( sumySq - ( sumy*sumy ) / n ) );
+    if(den > 0){
+      r = num / den;
+    }
+    return r*r;
+  } 
 >>
 
 INIT <<
-  // generate 500 case of poly-10 benchmark function
+  // generate 500 cases of poly-10 benchmark function
   int n = 500;
   variableNames = new string[] {"x1", "x2", "x3", "x4", "x5", "x6", "x7", "x8", "x9", "x10" };
@@ -116 +144 @@
   x = new double[n, 10];
   y = new double[n];
-  for(int row = 0; row < 500; row++) {
+  for(int row = 0; row < n; row++) {
     for(int col = 0; col < 10; col++) {
       x[row, col] = rand.NextDouble() * 2.0 - 1.0;
@@ -126 +154 @@
              x[row, 2] * x[row, 5] + x[row, 9];
   }
+ 
+  rows = System.Linq.Enumerable.Range(0, n).ToArray();
+  
+  // generate 100 random constants in [-10.0 .. 10.0[
+  randomConsts = Enumerable.Range(0, 100).Select(i => rand.NextDouble()*20.0 - 10.0).ToArray();
 >>
 
-/* non-terminals of the problem */
 NONTERMINALS
   Model<<int row, out double val>>.
@@ -137 +169 @@
   Division<<int row, out double val>>.
 
-/* terminals of the problem: random constants (ERC) and variables */
 TERMINALS
   ERC<<out double val>> 
     CONSTRAINTS
-          val IN RANGE <<-100>> .. <<100>>
+      val IN SET << randomConsts >>
   .
 
   Var<<out string varName>> 
     CONSTRAINTS
-          varName IN SET <<variableNames>>
-  .
-
-/* grammar rules for the problem with interleaved semantic actions */
+      varName IN SET << variableNames >>
+  .
+
 RULES
   Model<<int row, out double val>> =
@@ -160 +190 @@
     | Multiplication<<row, out val>>
     | Var<<out varName>>                                   SEM << val = GetValue(x, varName, row); >>
-    | ERC<<out val>>
+    /* | ERC<<out val>> */
   .
 
@@ -176 +206 @@
   .
 
-/* objective function */
-MAXIMIZE                                                   /* could also use the keyword MINIMIZE here */
+MAXIMIZE
   << 
-    var rows = System.Linq.Enumerable.Range(0, x.GetLength(0));
     var predicted = rows.Select(r => {
       double result;