数据科学和机器学习（第 05 部分）：决策树·综合运用

class="type">void CDecisionTree::MatrixClassify(class="type">class="kw">string &dataArr[],class="type">class="kw">string &Classes[], class="type">int cols)
{
  class="type">class="kw">string ClassifiedArr[];
  ArrayResize(ClassifiedArr,ArraySize(dataArr));

  class="type">int fill_start = class="num">0, fill_ends = class="num">0;
  class="type">int index = class="num">0;
  for (class="type">int i = class="num">0; i<ArraySize(Classes); i++)
  {
    class="type">int start = class="num">0;  class="type">int curr_col = class="num">0;
    for (class="type">int j = class="num">0; j<ArraySize(dataArr); j++)
    {
      curr_col++;

      if (Classes[i] == dataArr[j])
      {
        class=class="str">"cmt">//printf("Classes[%d] = %s dataArr[%d] = %s ",i,Classes[i],j,dataArr[j]);

        if (curr_col == class="num">1)
          fill_start =  j;

else
  {
    if (j>curr_col)
      fill_start = j - (curr_col-class="num">1);
    else fill_start = (curr_col-class="num">1) - j;

    fill_start  = fill_start;
    class=class="str">"cmt">//Print("j ",j," j-currcol ",j-(curr_col-class="num">1)," curr_col ",curr_col," columns ",cols," fill start ",fill_start );

  }

fill_ends = fill_start + cols;

class=class="str">"cmt">//printf("fillstart %d fillends %d j index = %d i = %d ",fill_start,fill_ends,j,i);
class=class="str">"cmt">//---
  class=class="str">"cmt">//if (ArraySize(ClassifiedArr) >= ArraySize(dataArr)) class="kw">break;
  class=class="str">"cmt">//Print("ArraySize Classified Arr ",ArraySize(ClassifiedArr)," dataArr size ",ArraySize(dataArr)," i ",i);


  for (class="type">int k=fill_start; k<fill_ends; k++)
    {
      index++;
      class=class="str">"cmt">//printf(" k %d index %d",k,index);
      class=class="str">"cmt">//printf("dataArr[%d] = %s index = %d",k,dataArr[k],index-class="num">1);
      ClassifiedArr[index-class="num">1] = dataArr[k];
    }

  if (index >= ArraySize(dataArr)) class="kw">break; class=class="str">"cmt">//might be infinite loop if this occurs
   }

  if (curr_col == cols) curr_col = class="num">0;
   }

  if (index >= ArraySize(dataArr)) class="kw">break; class=class="str">"cmt">//might be infinite loop if this occurs
   }

 ArrayCopy(dataArr,ClassifiedArr);
 ArrayFree(ClassifiedArr);
}
class=class="str">"cmt">// JG    class="num">0     class="num">13:class="num">47:class="num">20.574   TestScript(EURUSD,H1)   Classified matrix dataset

class="type">int zero_entropy_index = class="num">0;
class="type">bool zero_entropy = class="kw">false;
for (class="type">int e=class="num">0; e<ArraySize(P_EntropyArr); e++)
   if (P_EntropyArr[e] == class="num">0) { zero_entropy = true; zero_entropy_index=e; class="kw">break; }

if (zero_entropy) class=class="str">"cmt">//if there is zero in the Entropy Array 
  {
  MatrixRemoveRow(m_dataset,p_Classes[zero_entropy_index],cols);
  
  rows_total = ArraySize(m_dataset); class=class="str">"cmt">//New number of total rows from Array  
   if (m_debug)
    {
     printf("%s is A LEAF NODE its Rows have been removed from the dataset remaining Dataset is ..",p_Classes[zero_entropy_index]);
     ArrayPrint(DataColumnNames);
     MatrixPrint(m_dataset,cols,rows_total);
    }
  
  class=class="str">"cmt">//we also remove the entropy from the Array and its information everywhere else from the parent Node That we are going to build next
  

ArrayRemove(P_EntropyArr,zero_entropy_index,class="num">1);
ArrayRemove(p_Classes,zero_entropy_index,class="num">1);
ArrayRemove(p_ClassNumbers,zero_entropy_index,class="num">1);

class=class="str">"cmt">//---      REMOVING THE PARENT/ ROOT NODE FROM OUR DATASET
      MatrixRemoveColumn(m_dataset,max_gain,cols);
      
       class=class="str">"cmt">// After removing the columns assign the new values to these global variables
      
        cols = cols-class="num">1;   class=class="str">"cmt">// remove that one column that has been removed
        rows_total = rows_total - single_rowstotal; class=class="str">"cmt">//remove the size of one column rows
      
       class=class="str">"cmt">// we also remove the column from column names Array 
        ArrayRemove(DataColumnNames,max_gain,class="num">1);
class=class="str">"cmt">//---
        printf("Column %d removed from the Matrix, The remaining dataset is",max_gain+class="num">1);
        ArrayPrint(DataColumnNames);
        MatrixPrint(m_dataset,cols,rows_total);

class="type">void CDecisionTree::BuildTree(class="type">void)
{
   class="type">int ClassNumbers[];
   class="type">int max_gain = class="num">0;
   class="type">class="kw">double IGArr[];
   class=class="str">"cmt">//class="type">class="kw">double parent_entropy = Entropy(p_ClassNumbers,single_rowstotal);
   class="type">class="kw">string p_Classes[];    class=class="str">"cmt">//parent classes
   class="type">class="kw">double P_EntropyArr[]; class=class="str">"cmt">//Parent Entropy
   class="type">int p_ClassNumbers[]; class=class="str">"cmt">//parent/ Target variable class numbers
   GetClasses(TargetArr,m_DatasetClasses,p_ClassNumbers);
   ArrayResize(P_EntropyArr,class="num">1);
   P_EntropyArr[class="num">0] = Entropy(p_ClassNumbers,single_rowstotal);
class=class="str">"cmt">//--- temporary disposable arrays for parent node information
   class="type">class="kw">string TempP_Classes[];
   class="type">class="kw">double TempP_EntropyArr[];
   class="type">int    TempP_ClassNumbers[];
class=class="str">"cmt">//---
   if (m_debug) Print("Default Parent Entropy ",P_EntropyArr[class="num">0]);
   class="type">int cols = m_colschosen;
      for (class="type">int A =class="num">0; A<ArraySize(P_EntropyArr); A++)
      {
         printf("<<<<<<<<    Parent Entropy  %.5f  A = %d  >>>>>>>> ",P_EntropyArr[A],A);
         for (class="type">int i=class="num">0; i<cols-class="num">1; i++) class=class="str">"cmt">//we substract with one to remove the independent variable coumn
         {
            class="type">int rows = ArraySize(m_dataset)/cols;
            class="type">class="kw">string Arr[]; class=class="str">"cmt">//ArrayFor the current column
            class="type">class="kw">string ArrTarg[]; class=class="str">"cmt">//Array for the current target
            

ArrayResize(Arr,rows);
ArrayResize(ArrTarg,rows);

printf("  <<<<<   C O L U M N  %s  >>>>>  ",DataColumnNames[i]);
class="type">int index_target=cols-class="num">1;
for (class="type">int j=class="num">0; j<rows; j++) class=class="str">"cmt">//get column data and its target column
  {
   class="type">int index = i+j * cols;
   class=class="str">"cmt">//Print("index ",index);
   Arr[j] = m_dataset[index];

   class=class="str">"cmt">//printf("ArrTarg[%d] = %s m_dataset[%d] =%s ",j,ArrTarg[j],index_target,m_dataset[index_target]);
   ArrTarg[j] = m_dataset[index_target];

   class=class="str">"cmt">//printf("Arr[%d] = %s ArrTarg[%d] = %s ",j,Arr[j],j,ArrTarg[j]); 

   index_target += cols; class=class="str">"cmt">//the last index of all the columns
  } 

class=class="str">"cmt">//--- Finding the Entropy
class=class="str">"cmt">//The function to find the Entropy of samples in a given column inside its loop
class=class="str">"cmt">//then restores all the entropy into one array

class=class="str">"cmt">//--- Finding the Information Gain
class=class="str">"cmt">//The Function to find the information gain from the entropy array above
class=class="str">"cmt">//---
if (i == max_gain)
  {
   class=class="str">"cmt">//Get the maximum information gain of all the information gain in all columns then 

class=class="str">"cmt">//store it to the parent information gain
 }

 class=class="str">"cmt">//---

 ZeroMemory(ClassNumbers);
 ZeroMemory(SamplesNumbers);

 }

 class=class="str">"cmt">//---- Get the parent Entropy, class and class numbers
 class=class="str">"cmt">// here we store the obtained parent class from the information gain metric then we store them into a parent array
 ArrayCopy(p_Classes,TempP_Classes);
 ArrayCopy(P_EntropyArr,TempP_EntropyArr);
 ArrayCopy(p_ClassNumbers,TempP_ClassNumbers);

 class=class="str">"cmt">//---

 class="type">class="kw">string Node[class="num">1];
 Node[class="num">0] = DataColumnNames[max_gain];

 if (m_debug)
 printf("Parent Node will be %s with IG = %.5f",Node[class="num">0],IGArr[max_gain]);

 if (A == class="num">0)
 DrawTree(Node,"parent",A);
 DrawTree(p_Classes,"child",A);


 class=class="str">"cmt">//--- CLASSIFY THE MATRIX
 MatrixClassify(m_dataset,p_Classes,cols);

 class=class="str">"cmt">//--- Search if there is zero entropy in Array if there is any remove its data from the dataset

 if (P_EntropyArr[e] == class="num">0) { zero_entropy = true; zero_entropy_index=e; class="kw">break; }

 if (zero_entropy) class=class="str">"cmt">//if there is zero in the Entropy Array 
 {
 MatrixRemoveRow(m_dataset,p_Classes[zero_entropy_index],cols);

 rows_total = ArraySize(m_dataset); class=class="str">"cmt">//New number of total rows from Array

 class=class="str">"cmt">//we also remove the entropy from the Array and its information everywhere else from the parent Node That we are going to build next

      ArrayRemove(P_EntropyArr,zero_entropy_index,class="num">1);
      ArrayRemove(p_Classes,zero_entropy_index,class="num">1);
      ArrayRemove(p_ClassNumbers,zero_entropy_index,class="num">1);
      }
      
      if (m_debug)  Print("rows total ",rows_total," ",p_Classes[zero_entropy_index]," ",p_ClassNumbers[zero_entropy_index]);
      
class=class="str">"cmt">//---    REMOVING THE PARENT/ ROOT NODE FROM OUR DATASET
      MatrixRemoveColumn(m_dataset,max_gain,cols);
      
       class=class="str">"cmt">// After removing the columns assing the new values to these global variables
      
      cols = cols-class="num">1;   class=class="str">"cmt">// remove that one column that has been removed
      rows_total = rows_total - single_rowstotal; class=class="str">"cmt">//remove the size of one column rows
      
       class=class="str">"cmt">// we also remove the column from column names Array 
      ArrayRemove(DataColumnNames,max_gain,class="num">1);
      
class=class="str">"cmt">//---
      
   }
   
}