Changeset 14351 for branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic

Timestamp:

10/23/16 09:44:29 (8 years ago)

Author:

gkronber

Message:

#2650: merged r14332:14350 from trunk to branch

Location:

branches/symbreg-factors-2650

Files:

• . (modified) (1 prop)
• HeuristicLab.Problems.DataAnalysis.Symbolic (modified) (1 prop)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Formatters/InfixExpressionFormatter.cs (modified) (4 diffs)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Importer/InfixExpressionParser.cs (modified) (11 diffs)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionCompiledTreeInterpreter.cs (modified) (1 diff)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionTreeInterpreter.cs (modified) (1 diff)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionTreeLinearInterpreter.cs (modified) (3 diffs)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Symbols/VariableCondition.cs (modified) (1 diff)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Symbols/VariableConditionTreeNode.cs (modified) (1 diff)

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Formatters/InfixExpressionFormatter.cs

@@ -81 +81 @@
           }
           strBuilder.Append(")");
+        } else {
+          // function with multiple arguments
+          strBuilder.Append(token).Append("(");
+          FormatRecursively(node.Subtrees.First(), strBuilder);
+          foreach (var subtree in node.Subtrees.Skip(1)) {
+            strBuilder.Append(", ");
+            FormatRecursively(subtree, strBuilder);
+          }
+          strBuilder.Append(")");
         }
       } else if (node.SubtreeCount == 1) {
@@ -94 +103 @@
           FormatRecursively(node.GetSubtree(0), strBuilder);
         } else {
-          // function
+          // function with only one argument
           strBuilder.Append(token).Append("(");
           FormatRecursively(node.GetSubtree(0), strBuilder);
@@ -101 +110 @@
       } else {
         // no subtrees
-        if (node.Symbol is Variable) {
+        if (node.Symbol is LaggedVariable) {
+          var varNode = node as LaggedVariableTreeNode;
+          if (!varNode.Weight.IsAlmost(1.0)) {
+            strBuilder.Append("(");
+            strBuilder.AppendFormat(CultureInfo.InvariantCulture, "{0}", varNode.Weight);
+            strBuilder.Append("*");
+          }
+          strBuilder.Append("LAG(");
+          if (varNode.VariableName.Contains("'")) {
+            strBuilder.AppendFormat("\"{0}\"", varNode.VariableName);
+          } else {
+            strBuilder.AppendFormat("'{0}'", varNode.VariableName);
+          }
+          strBuilder.Append(", ")
+            .AppendFormat(CultureInfo.InvariantCulture, "{0}", varNode.Lag)
+            .Append(")");
+        } else if (node.Symbol is Variable) {
           var varNode = node as VariableTreeNode;
           if (!varNode.Weight.IsAlmost(1.0)) {
@@ -144 +169 @@
             strBuilder.AppendFormat(CultureInfo.InvariantCulture, "{0}", constNode.Value);
           else
-            strBuilder.AppendFormat(CultureInfo.InvariantCulture, "({0})", constNode.Value);     // (-1)
+            strBuilder.AppendFormat(CultureInfo.InvariantCulture, "({0})", constNode.Value); // (-1
         }
       }

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Importer/InfixExpressionParser.cs

@@ -26 +26 @@
 using System.Text;
 using HeuristicLab.Collections;
+using HeuristicLab.Common;
 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
 
@@ -48 +49 @@
   /// </summary>
   public sealed class InfixExpressionParser {
-    private enum TokenType { Operator, Identifier, Number, LeftPar, RightPar, Eq, End, NA };
+    private enum TokenType { Operator, Identifier, Number, LeftPar, RightPar, Comma, Eq, End, NA };
     private class Token {
       internal double doubleVal;
@@ -114 +115 @@
         { "MEAN", new Average()},
         { "IF", new IfThenElse()},
-        { ">", new GreaterThan()},
-        { "<", new LessThan()},
+        { "GT", new GreaterThan()},
+        { "LT", new LessThan()},
         { "AND", new And()},
         { "OR", new Or()},
@@ -121 +122 @@
         { "XOR", new Xor()},
         { "DIFF", new Derivative()},
+        { "LAG", new LaggedVariable() },
       };
 
@@ -150 +152 @@
         }
         if (char.IsDigit(str[pos])) {
-          // read number (=> read until white space or operator)
+          // read number (=> read until white space or operator or comma)
           var sb = new StringBuilder();
           sb.Append(str[pos]);
@@ -159 +161 @@
             && str[pos] != '*'
             && str[pos] != '/'
-            && str[pos] != ')') {
+            && str[pos] != ')'
+            && str[pos] != ',') {
             sb.Append(str[pos]);
             pos++;
@@ -226 +229 @@
           pos++;
           yield return new Token { TokenType = TokenType.Eq, strVal = "=" };
+        } else if (str[pos] == ',') {
+          pos++;
+          yield return new Token { TokenType = TokenType.Comma, strVal = "," };
         } else {
           throw new ArgumentException("Invalid character: " + str[pos]);
@@ -232 +238 @@
     }
 
+    // S       = Expr EOF
+    // Expr    = ['-' | '+'] Term { '+' Term | '-' Term }
+    // Term    = Fact { '*' Fact | '/' Fact }
+    // Fact    = '(' Expr ')' | funcId '(' ArgList ')' | varId | number
+    // ArgList = Expr { ',' Expr }
     private ISymbolicExpressionTreeNode ParseS(Queue<Token> tokens) {
       var expr = ParseExpr(tokens);
@@ -339 +350 @@
     }
 
-    // Fact = '(' Expr ')' | funcId '(' Expr ')' | varId [ = valId ] | number
+    // Fact = '(' Expr ')' | 'LAG' '(' varId ',' ['+' | '-' ] number ')' | funcId '(' Expr ')' | varId [ = valId ] | number
     private ISymbolicExpressionTreeNode ParseFact(Queue<Token> tokens) {
       var next = tokens.Peek();
@@ -359 +370 @@
           if (lPar.TokenType != TokenType.LeftPar)
             throw new ArgumentException("expected (");
-          var expr = ParseExpr(tokens);
+
+          // handle 'lag' specifically
+          if (funcNode.Symbol is LaggedVariable) {
+            var varId = tokens.Dequeue();
+            if (varId.TokenType != TokenType.Identifier) throw new ArgumentException("Identifier expected. Format for lagged variables: \"lag(x, -1)\"");
+            var comma = tokens.Dequeue();
+            if (comma.TokenType != TokenType.Comma) throw new ArgumentException("',' expected, Format for lagged variables: \"lag(x, -1)\"");
+            double sign = 1.0;
+            if (tokens.Peek().strVal == "+" || tokens.Peek().strVal == "-") {
+              // read sign
+              var signTok = tokens.Dequeue();
+              if (signTok.strVal == "-") sign = -1.0;
+            }
+            var lagToken = tokens.Dequeue();
+            if (lagToken.TokenType != TokenType.Number) throw new ArgumentException("Number expected, Format for lagged variables: \"lag(x, -1)\"");
+            if (!lagToken.doubleVal.IsAlmost(Math.Round(lagToken.doubleVal)))
+              throw new ArgumentException("Time lags must be integer values");
+            var laggedVarNode = funcNode as LaggedVariableTreeNode;
+            laggedVarNode.VariableName = varId.strVal;
+            laggedVarNode.Lag = (int)Math.Round(sign * lagToken.doubleVal);
+            laggedVarNode.Weight = 1.0;
+          } else {
+            // functions
+            var args = ParseArgList(tokens);
+            // check number of arguments
+            if (funcNode.Symbol.MinimumArity > args.Length || funcNode.Symbol.MaximumArity < args.Length) {
+              throw new ArgumentException(string.Format("Symbol {0} requires between {1} and  {2} arguments.", funcId,
+                funcNode.Symbol.MinimumArity, funcNode.Symbol.MaximumArity));
+            }
+            foreach (var arg in args) funcNode.AddSubtree(arg);
+          }
+
           var rPar = tokens.Dequeue();
           if (rPar.TokenType != TokenType.RightPar)
             throw new ArgumentException("expected )");
 
-          funcNode.AddSubtree(expr);
           return funcNode;
         } else {
@@ -396 +437 @@
     }
 
+    // ArgList = Expr { ',' Expr }
+    private ISymbolicExpressionTreeNode[] ParseArgList(Queue<Token> tokens) {
+      var exprList = new List<ISymbolicExpressionTreeNode>();
+      exprList.Add(ParseExpr(tokens));
+      while (tokens.Peek().TokenType != TokenType.RightPar) {
+        var comma = tokens.Dequeue();
+        if (comma.TokenType != TokenType.Comma) throw new ArgumentException("expected ',' ");
+        exprList.Add(ParseExpr(tokens));
+      }
+      return exprList.ToArray();
+    }
   }
 }

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionCompiledTreeInterpreter.cs

@@ -614 +614 @@
         case OpCodes.VariableCondition: {
             var variableConditionTreeNode = (VariableConditionTreeNode)node;
+            if (variableConditionTreeNode.Symbol.IgnoreSlope) throw new NotSupportedException("Strict variable conditionals are not supported");
             var variableName = variableConditionTreeNode.VariableName;
             var indexExpr = Expression.Constant(variableIndices[variableName]);

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionTreeInterpreter.cs

@@ -488 +488 @@
             if (row < 0 || row >= dataset.Rows) return double.NaN;
             var variableConditionTreeNode = (VariableConditionTreeNode)currentInstr.dynamicNode;
-            double variableValue = ((IList<double>)currentInstr.data)[row];
-            double x = variableValue - variableConditionTreeNode.Threshold;
-            double p = 1 / (1 + Math.Exp(-variableConditionTreeNode.Slope * x));
-
-            double trueBranch = Evaluate(dataset, ref row, state);
-            double falseBranch = Evaluate(dataset, ref row, state);
-
-            return trueBranch * p + falseBranch * (1 - p);
+            if (!variableConditionTreeNode.Symbol.IgnoreSlope) {
+              double variableValue = ((IList<double>)currentInstr.data)[row];
+              double x = variableValue - variableConditionTreeNode.Threshold;
+              double p = 1 / (1 + Math.Exp(-variableConditionTreeNode.Slope * x));
+
+              double trueBranch = Evaluate(dataset, ref row, state);
+              double falseBranch = Evaluate(dataset, ref row, state);
+
+              return trueBranch * p + falseBranch * (1 - p);
+            } else {
+              // strict threshold
+              double variableValue = ((IList<double>)currentInstr.data)[row];
+              if (variableValue <= variableConditionTreeNode.Threshold) {
+                var left = Evaluate(dataset, ref row, state);
+                state.SkipInstructions();
+                return left;
+              } else {
+                state.SkipInstructions();
+                return Evaluate(dataset, ref row, state);
+              }
+            }
           }
         default:

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Interpreter/SymbolicDataAnalysisExpressionTreeLinearInterpreter.cs

@@ -126 +126 @@
     private readonly object syncRoot = new object();
     public IEnumerable<double> GetSymbolicExpressionTreeValues(ISymbolicExpressionTree tree, IDataset dataset, IEnumerable<int> rows) {
+      if (!rows.Any()) return Enumerable.Empty<double>();
       if (CheckExpressionsWithIntervalArithmetic)
         throw new NotSupportedException("Interval arithmetic is not yet supported in the symbolic data analysis interpreter.");
@@ -171 +172 @@
           if (row < 0 || row >= dataset.Rows) instr.value = double.NaN;
           var variableConditionTreeNode = (VariableConditionTreeNode)instr.dynamicNode;
-          double variableValue = ((IList<double>)instr.data)[row];
-          double x = variableValue - variableConditionTreeNode.Threshold;
-          double p = 1 / (1 + Math.Exp(-variableConditionTreeNode.Slope * x));
-
-          double trueBranch = code[instr.childIndex].value;
-          double falseBranch = code[instr.childIndex + 1].value;
-
-          instr.value = trueBranch * p + falseBranch * (1 - p);
+          if (!variableConditionTreeNode.Symbol.IgnoreSlope) {
+            double variableValue = ((IList<double>)instr.data)[row];
+            double x = variableValue - variableConditionTreeNode.Threshold;
+            double p = 1 / (1 + Math.Exp(-variableConditionTreeNode.Slope * x));
+
+            double trueBranch = code[instr.childIndex].value;
+            double falseBranch = code[instr.childIndex + 1].value;
+
+            instr.value = trueBranch * p + falseBranch * (1 - p);
+          } else {
+            double variableValue = ((IList<double>)instr.data)[row];
+            if (variableValue <= variableConditionTreeNode.Threshold) {
+              instr.value = code[instr.childIndex].value;
+            } else {
+              instr.value = code[instr.childIndex + 1].value;
+            }
+          }
         } else if (instr.opCode == OpCodes.Add) {
           double s = code[instr.childIndex].value;
@@ -433 +443 @@
               for (int j = 1; j != seq.Length; ++j)
                 seq[j].skip = true;
-            }
-            break;
+              break;
+            }
         }
         #endregion

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Symbols/VariableCondition.cs

@@ -149 +149 @@
         }
       }
+    }
+
+    /// <summary>
+    /// Flag to indicate if the interpreter should ignore the slope parameter (introduced for representation of expression trees)
+    /// </summary>
+    [Storable]
+    private bool ignoreSlope;
+    public bool IgnoreSlope {
+      get { return ignoreSlope; }
+      set { ignoreSlope = value; }
     }
 

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Symbols/VariableConditionTreeNode.cs

@@ -96 +96 @@
 
     public override string ToString() {
-      if (slope.IsAlmost(0.0))
+      if (slope.IsAlmost(0.0) || Symbol.IgnoreSlope) {
+        return variableName + " < " + threshold.ToString("E4");
+      } else {
         return variableName + " > " + threshold.ToString("E4") + Environment.NewLine +
-          "slope: " + slope.ToString("E4");
-      else
-        return variableName + " > " + threshold.ToString("E4");
+               "slope: " + slope.ToString("E4");
+      }
     }
   }