Java机器学习库ML之七分类预测输出概率值

场景：一般分类预测直接输出类别标记，不过有些情况需要输出对应类别的概率值，比如判定为正例的概率是0.6，而判定为负例的概率是0.3，那自然标记为正例，这里就是看ML用classDistribution输出各类别的概率值。参考代码如下：

/*** This file is part of the Java Machine Learning Library* * The Java Machine Learning Library is free software; you can redistribute it and/or modify* it under the terms of the GNU General Public License as published by* the Free Software Foundation; either version 2 of the License, or* (at your option) any later version.* * The Java Machine Learning Library is distributed in the hope that it will be useful,* but WITHOUT ANY WARRANTY; without even the implied warranty of* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the* GNU General Public License for more details.* * You should have received a copy of the GNU General Public License* along with the Java Machine Learning Library; if not, write to the Free Software* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA* * Copyright (c) 2006-2012, Thomas Abeel* * Project: http://java-ml.sourceforge.net/* */
package com.gddx;import java.io.File;
import java.util.Map;
import java.util.Random;import be.abeel.util.Pair;
import net.sf.javaml.classification.Classifier;
import net.sf.javaml.classification.tree.RandomForest;
import net.sf.javaml.core.Dataset;
import net.sf.javaml.core.DenseInstance;
import net.sf.javaml.core.Instance;
import net.sf.javaml.sampling.Sampling;
import net.sf.javaml.tools.data.FileHandler;/*** Tutorial for the random forest classifier.* * @author Thomas Abeel* */
public class TutorialRandomForest {/*** Shows the default usage of the random forest algorithm.*/public static void main(String[] args) throws Exception {/* Load a data set */Dataset ori_data = FileHandler.loadDataset(new File("D:\\tmp\\javaml-0.1.7-src\\UCI-small\\iris\\iris.data"), 4, ",");Sampling s = Sampling.SubSampling;Pair<Dataset, Dataset> sam_data = s.sample(ori_data, (int) (ori_data.size() * 0.8));/** Contruct a RF classifier that uses 5 neighbors to make a decision.*/Classifier rf = new RandomForest(50, false, 3, new Random());rf.buildClassifier(sam_data.x());//80%样本训练/* 输出预测的类别概率 */for(Instance inst:sam_data.y()){  //20%样本验证Map<Object,Double> mprob=rf.classDistribution(inst);//输出类别的概率，[0,1]System.out.println(mprob);       }}}

执行结果：

{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=0.7400000000000003, Iris-versicolor=0.25999999999999995}
{Iris-virginica=0.0, Iris-setosa=1.0000000000000004, Iris-versicolor=0.0}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=0.08, Iris-setosa=0.0, Iris-versicolor=0.9200000000000005}
{Iris-virginica=0.46000000000000013, Iris-setosa=0.0, Iris-versicolor=0.5400000000000001}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=0.0, Iris-setosa=0.0, Iris-versicolor=1.0000000000000004}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=0.32, Iris-setosa=0.0, Iris-versicolor=0.6800000000000003}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=1.0000000000000004, Iris-setosa=0.0, Iris-versicolor=0.0}
{Iris-virginica=0.9000000000000005, Iris-setosa=0.0, Iris-versicolor=0.1}

Java机器学习库ML之七分类预测输出概率值相关推荐

Java机器学习库ML之六关于模型迭代训练的思考
我遇到的场景是:样本集有5000万条,接近5个G,那么这样的样本集一次导入训练,我放着一天一夜都没跑出结果,机器性能还特别好,是64位linux有128G内存. 针对这样的情况,我想到的是两种思路: ...
Java机器学习库ML之一Dataset和Instance
Java机器学习库ML官网:http://java-ml.sourceforge.net/ 对于一个机器学习库来说,最基础就是数据处理能力,ml库给了dataset和instance两个类,datas ...
Java机器学习库ML之四模型训练和预测示例
基于ML库机器学习的步骤: 1)样本数据导入: 2)样本数据特征抽取和特征值处理(结合模型需要归一化或离散化):这里本文没有做处理,特征选择和特征值处理本身就很大: 3)样本集划分训练集和验证集: 4 ...
Java机器学习库ML之五样本不均衡
样本不均衡的问题是指训练集中类别之间的样本数据量比例失衡.有研究表明,部分学习任务,在类别出现1∶35比例时就会使某些分类学习方法无效,甚至1∶10比例也会.样本不均衡导致的问题有: 1)少数类所包含 ...
Java机器学习库ML之二Feature Selection（特征选择）
机器学习中训练模型的前提必备工作就是特征选择,起到降维和降低计算开销的目的,当然在获取尽可能小的特征子集的同时,应不显著降低分类精度.不影响类分布.保持特征子集的稳定适应性强等. ML库提供了特征选择 ...
Java机器学习库ML之八关于模型迭代训练的试验
前文提到因为数据集过大,只能拆分然后依次迭代训练,实验发现对结果有所偏差,参考代码如下: package com.vip;import java.io.File; import java.util.M ...
Java机器学习库ML之十一线性SVM
线性SVM的原理就不多说了,最强大的就是libsvm库(ml库也是用这个),参考:http://blog.csdn.net/fjssharpsword/article/details/53883340 ...
Java机器学习库ML之十模型选择准则AIC和BIC
学习任务所建立的模型多数是参数估计并采用似然函数作为目标函数,当训练数据足够多时,可以不断提高模型精度,但是以提高模型复杂度为代价的,同时也带来一个机器学习中非常普遍的问题--过拟合.模型选择问题是在 ...
Java机器学习库ML之九交叉验证法(Cross Validation)
交叉验证(Cross Validation,CV)是用来验证分类器的性能一种统计分析方法,基本思想是把在某种意义下将原始数据(dataset)进行分组,一部分做为训练集(train set),另一部分 ...

Java机器学习库ML之七分类预测输出概率值

Java机器学习库ML之七分类预测输出概率值相关推荐

最新文章

热门文章