隨機森林 RandomForest java
隨機森林是由多棵樹組成的分類或迴歸方法。主要思想來源於Bagging演算法,Bagging技術思想主要是給定一弱分類器及訓練集,讓該學習演算法訓練多輪,每輪的訓練集由原始訓練集中有放回的隨機抽取,大小一般跟原始訓練集相當,這樣依次訓練多個弱分類器,最終的分類由這些弱分類器組合,對於分類問題一般採用多數投票法,對於迴歸問題一般採用簡單平均法。隨機森林在bagging的基礎上,每個弱分類器都是決策樹,決策樹的生成過程中中,在屬性的選擇上增加了依一定概率選擇屬性,在這些屬性中選擇最佳屬性及分割點,傳統做法一般是全部屬性中去選擇最佳屬性,這樣隨機森林有了樣本選擇的隨機性,屬性選擇的隨機性,這樣一來增加了每個分類器的差異性、不穩定性及一定程度上避免每個分類器的過擬合(一般決策樹有過擬合現象),由此組合分類器增加了最終的泛化能力。下面是程式碼的簡單實現
/**
* 隨機森林 迴歸問題
* @author ysh 1208706282
*
*/
public class RandomForest {
List<Sample> mSamples;
List<Cart> mCarts;
double mFeatureRate;
int mMaxDepth;
int mMinLeaf;
Random mRandom;
/**
* 載入資料 迴歸樹
* @param path
* @param regex
* @throws Exception
*/
public void loadData(String path,String regex) throws Exception{
mSamples = new ArrayList<Sample>();
BufferedReader reader = new BufferedReader(new FileReader(path));
String line = null;
String splits[] = null;
Sample sample = null;
while(null != (line=reader.readLine())){
splits = line.split(regex);
sample = new Sample();
sample.label = Double.valueOf(splits[0]);
sample.feature = new ArrayList<Double>(splits.length-1);
for(int i=0;i<splits.length-1;i++){
sample.feature.add(new Double(splits[i+1]));
}
mSamples.add(sample);
}
reader.close();
}
public void train(int iters){
mCarts = new ArrayList<Cart>(iters);
Cart cart = null;
for(int iter=0;iter<iters;iter++){
cart = new Cart();
cart.mFeatureRate = mFeatureRate;
cart.mMaxDepth = mMaxDepth;
cart.mMinLeaf = mMinLeaf;
cart.mRandom = mRandom;
List<Sample> s = new ArrayList<Sample>(mSamples.size());
for(int i=0;i<mSamples.size();i++){
s.add(mSamples.get(cart.mRandom.nextInt(mSamples.size())));
}
cart.setData(s);
cart.train();
mCarts.add(cart);
System.out.println("iter: "+iter);
s = null;
}
}
/**
* 迴歸問題簡單平均法 分類問題多數投票法
* @param sample
* @return
*/
public double classify(Sample sample){
double val = 0;
for(Cart cart:mCarts){
val += cart.classify(sample);
}
return val/mCarts.size();
}
/**
* @param args
* @throws Exception
*/
public static void main(String[] args) throws Exception {
// TODO Auto-generated method stub
RandomForest forest = new RandomForest();
forest.loadData("F:/2016-contest/20161001/train_data_1.csv", ",");
forest.mFeatureRate = 0.8;
forest.mMaxDepth = 3;
forest.mMinLeaf = 1;
forest.mRandom = new Random();
forest.mRandom.setSeed(100);
forest.train(100);
List<Sample> samples = Cart.loadTestData("F:/2016-contest/20161001/valid_data_1.csv", true, ",");
double sum = 0;
for(Sample s:samples){
double val = forest.classify(s);
sum += (val-s.label)*(val-s.label);
System.out.println(val+" "+s.label);
}
System.out.println(sum/samples.size()+" "+sum);
System.out.println(System.currentTimeMillis());
}
}