first runnable with trainData given by our teacher, just 2 trees, each 22s to run 7 train data and 3 test data. Using all the features at nodes split.

Author: lzddzh

LoadData.cpp

@@ -61,7 +61,6 @@ void LoadData::convertRawDataToExamples() {
             if(y_u<cropingWidth&&x_v>=0)
             {
                 t.x[j]= rawData[i].terms[x*cropingWidth+y_u]-rawData[i].terms[x_v*cropingWidth+y];
-                cout<<t.x[j]<<endl;
             }
             //out of bound , during training we don't use this feature if any picture is out of bound case.
             else

README.md

@@ -33,7 +33,7 @@ ExampleInput/  *Stores some small examples of input data file.
 * g++  * The C++ compiler, or you can use anything equvalent.
 * Armadillo-7.400.2   * A linear algebra C++ open library.
 
-**How to installi the prerequests?**
+**How to install the prerequests?**
 First download the `armdillo-7.400.2.zip` from [this link](http://pan.baidu.com/s/1o7Um4Sa), then uppack it. 
 Alternatively, you could also refers to [official website of Armdillo](http://arma.sourceforge.net/)
 
@@ -87,7 +87,7 @@ for now, just choose `debug mode`.
 This program does not input from the raw files like .shand, .skdepth or skdepth.cropping.
 Instead this program takes a single '.csv' file as its input, which stores all the training examples.
 
-An example input .csv file is at `HandPoseEstimation/ExampleInput/`
+The training Data in .csv format can be downloaded at: [this link](https://pan.baidu.com/s/1dFNui1j)
 where `value0, value1, ...` is the pixel value of our depth image, so there should be `90*60` values,and the `label0, label1, ...` is the hand joints postion value, so there should be `20*3` labels.
 
 In both `debug` and `real-time` mode of this program at running time, the program will output two files

Tree.cpp

@@ -114,8 +114,8 @@ example averageLabelValueOfExamples(const vector<example> &E) {
 }
 // By Tony, Sep 20.
 vector<feature> randomChooseFeatures() {
-	int num = int(sqrt(feaNum));
-	//int num = 100;
+	//int num = int(sqrt(feaNum));
+	int num = feaNum;
 	bool used[feaNum] = { 0 };
 	vector<feature> feats;
 	for (int i = 0; i < num; i++) {
@@ -152,7 +152,7 @@ double H(const vector<example> &E) {
     //Calculate the determinant of covS.
     double detCS = arma::det(covS);
         // FOR DEBUG ONLY:
-    cout << "detcs = " << detCS << endl;
+    //cout << "detcs = " << detCS << endl;
     // log() is the function 'ln()'
     // M_PI = 3.14159265358979323846
     double c = (labelNum / 2) * (1 + log(2 * M_PI));
@@ -178,8 +178,8 @@ inline double infoGain(const vector<example> &E, const vector<example> &upSet, c
         return INF - downSet.size() * H(downSet);  
     if (H(downSet) == -1 * numeric_limits<double>::infinity())
         return INF - upSet.size() * H(upSet);  
-    cout << "size of UPset: " << upSet.size() << "size of DownSet: " << downSet.size() << endl;
-    cout << "H(E) of UpSet: " << H(upSet) << "  H(E) of DownSet: " << H(downSet) << endl;
+    //cout << "size of UPset: " << upSet.size() << "size of DownSet: " << downSet.size() << endl;
+    //cout << "H(E) of UpSet: " << H(upSet) << "  H(E) of DownSet: " << H(downSet) << endl;
     return H(E) - ((upSet.size() / E.size()) * H(upSet) + (downSet.size() / E.size() * H(downSet)));
 }
 // By Tony. Sep 20.
@@ -193,8 +193,8 @@ feature chooseBestFeature(const vector<feature> &feats, vector<example> &E, vect
 			if (E[j].x[i] > max) max = E[j].x[i];
 			if (E[j].x[i] < min) min = E[j].x[i];
 		}
-        cout << " i = " << i << "max = " << max << "min = " << min << endl;
-        if (0.01 > max - min || 0.01 > min - max) continue;
+        //cout << " i = " << i << "max = " << max << "min = " << min << endl;
+        if (0.01 > max - min && 0.01 > min - max) continue;
         // Try a number of split points.
 		double split[splitPointsNum], step = (max - min) / (splitPointsNum + 1);
 		for (int j = 0; j < splitPointsNum; j++) split[j] = min + step * (j + 1); 
@@ -212,7 +212,7 @@ feature chooseBestFeature(const vector<feature> &feats, vector<example> &E, vect
 			}
 			double nu = infoGain(E, upSet, downSet);
             // For debug only.
-            cout << "nu = " << nu << endl;
+            //cout << "nu = " << nu << endl;
             // Get the best feature index. and the split point value. 
 			if (maxInfoGain < nu) {
 				maxInfoGain = nu;
@@ -228,6 +228,7 @@ feature chooseBestFeature(const vector<feature> &feats, vector<example> &E, vect
         cout << "Exit program.." << endl;
         exit(0);
     }
+    cout << "bestFeature = " << bestFeature << endl;
 	for (int k = 0; k < E.size(); k++) {
 		if (E[k].x[bestFeature] <= sP) {
 			upSet.push_back(E[k]);