OneVsRest (OneVsRest)

Java 类名：com.alibaba.alink.pipeline.classification.OneVsRest

Python 类名：OneVsRest

功能介绍

本组件用One VS Rest策略进行多分类。

参数说明

名称	中文名称	描述	类型	是否必须？	取值范围	默认值
numClass	类别数	多分类的类别数，必选	Integer	✓
predictionCol	预测结果列名	预测结果列名	String	✓
modelFilePath	模型的文件路径	模型的文件路径	String			null
overwriteSink	是否覆写已有数据	是否覆写已有数据	Boolean			false
predictionDetailCol	预测详细信息列名	预测详细信息列名	String
reservedCols	算法保留列名	算法保留列	String[]			null
numThreads	组件多线程线程个数	组件多线程线程个数	Integer			1

代码示例

Python 代码

from pyalink.alink import *

import pandas as pd

useLocalEnv(1)

URL = "https://alink-test-data.oss-cn-hangzhou.aliyuncs.com/iris.csv";
SCHEMA_STR = "sepal_length double, sepal_width double, petal_length double, petal_width double, category string";
data = CsvSourceBatchOp().setFilePath(URL).setSchemaStr(SCHEMA_STR)

lr = LogisticRegression() \
    .setFeatureCols(["sepal_length", "sepal_width", "petal_length", "petal_width"]) \
    .setLabelCol("category") \
    .setPredictionCol("pred_result") \
    .setMaxIter(100)

oneVsRest = OneVsRest().setClassifier(lr).setNumClass(3)
model = oneVsRest.fit(data)
model.setPredictionCol("pred_result").setPredictionDetailCol("pred_detail")
model.transform(data).print()

Java 代码

import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.source.CsvSourceBatchOp;
import com.alibaba.alink.pipeline.classification.LogisticRegression;
import com.alibaba.alink.pipeline.classification.OneVsRest;
import com.alibaba.alink.pipeline.classification.OneVsRestModel;
import org.junit.Test;

public class OneVsRestTest {
	@Test
	public void testOneVsRest() throws Exception {
		String URL = "https://alink-test-data.oss-cn-hangzhou.aliyuncs.com/iris.csv";
		String SCHEMA_STR
			= "sepal_length double, sepal_width double, petal_length double, petal_width double, category string";
		BatchOperator <?> data = new CsvSourceBatchOp().setFilePath(URL).setSchemaStr(SCHEMA_STR);
		LogisticRegression lr = new LogisticRegression()
			.setFeatureCols("sepal_length", "sepal_width", "petal_length", "petal_width")
			.setLabelCol("category")
			.setPredictionCol("pred_result")
			.setMaxIter(100);
		OneVsRest oneVsRest = new OneVsRest().setClassifier(lr).setNumClass(3);
		OneVsRestModel model = oneVsRest.fit(data);
		model.setPredictionCol("pred_result").setPredictionDetailCol("pred_detail");
		model.transform(data).print();
	}
}

运行结果

sepal_length	sepal_width	petal_length	petal_width	category	pred_result	pred_detail
6.7000	3.1000	4.4000	1.4000	Iris-versicolor	Iris-versicolor	{“Iris-versicolor”:0.9999890601537083,“Iris-virginica”:1.0939842119301402E-5,“Iris-setosa”:4.1724971938972156E-12}
5.4000	3.0000	4.5000	1.5000	Iris-versicolor	Iris-versicolor	{“Iris-versicolor”:0.9939699721610056,“Iris-virginica”:0.006030026623291463,“Iris-setosa”:1.2157029667713158E-9}
5.4000	3.9000	1.7000	0.4000	Iris-setosa	Iris-setosa	{“Iris-versicolor”:0.02236524089333592,“Iris-virginica”:0.0,“Iris-setosa”:0.9776347591066641}
5.0000	3.4000	1.6000	0.4000	Iris-setosa	Iris-setosa	{“Iris-versicolor”:0.07720412400682967,“Iris-virginica”:0.0,“Iris-setosa”:0.9227958759931704}
5.6000	3.0000	4.5000	1.5000	Iris-versicolor	Iris-versicolor	{“Iris-versicolor”:0.9961816818708689,“Iris-virginica”:0.003818317908880254,“Iris-setosa”:2.2025091271297693E-10}
…	…	…	…	…	…	…