Changeset 12635

Timestamp:

07/07/15 12:37:11 (9 years ago)

Author:

gkronber

Message:

#2261: marked potential future efficiency improvements as identified through profiling

Location:

branches/GBT-trunkintegration/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees

Files:

• GradientBoostedTreesAlgorithmStatic.cs (modified) (1 diff)
• LossFunctions/AbsoluteErrorLoss.cs (modified) (1 diff)
• LossFunctions/RelativeErrorLoss.cs (modified) (1 diff)
• RegressionTreeModel.cs (modified) (2 diffs)

branches/GBT-trunkintegration/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/GradientBoostedTreesAlgorithmStatic.cs

@@ -180 +180 @@
 
       int i = 0;
+      // TODO: slow because of multiple calls to GetDoubleValue for each row index 
       foreach (var pred in tree.GetEstimatedValues(problemData.Dataset, problemData.TrainingIndices)) {
         yPred[i] = yPred[i] + nu * pred;

branches/GBT-trunkintegration/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/LossFunctions/AbsoluteErrorLoss.cs

@@ -84 +84 @@
           res[i - startIdx] = targetArr[row] - predArr[row];
         }
-        return res.Median();
+        return res.Median(); // TODO: improve efficiency
       };
       return lineSearch;

branches/GBT-trunkintegration/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/LossFunctions/RelativeErrorLoss.cs

@@ -100 +100 @@
             totalWeight += w;
           }
+          // TODO: improve efficiency (find median without sort)
           res_w.StableSort((a, b) => Math.Sign(a.Key - b.Key));
 

branches/GBT-trunkintegration/HeuristicLab.Algorithms.DataAnalysis/3.4/GradientBoostedTrees/RegressionTreeModel.cs

@@ -36 +36 @@
     public struct TreeNode {
       public readonly static string NO_VARIABLE = string.Empty;
-      public string VarName { get; set; } // name of the variable for splitting or -1 if terminal node
+      public string VarName { get; set; } // name of the variable for splitting or NO_VARIABLE if terminal node
       public double Val { get; set; } // threshold
       public int LeftIdx { get; set; }
@@ -66 +66 @@
       if (node.VarName == TreeNode.NO_VARIABLE)
         return node.Val;
+      // TODO: many calls to GetDoubleValue are slow because of the dictionary lookup in Dataset (see ticket #2417)
       else if (ds.GetDoubleValue(node.VarName, row) <= node.Val)
         return GetPredictionForRow(t, node.LeftIdx, ds, row);