岭回归 (RidgeRegression)

Java 类名：com.alibaba.alink.pipeline.regression.RidgeRegression
Python 类名：RidgeRegression

功能介绍

岭回归(Ridge regression)算法是一种经典的回归算法。岭回归组件支持稀疏、稠密两种数据格式，并且支持带权重样本训练。

算法原理

岭回归是一种专用于共线性数据分析的有偏估计回归方法，实质上是一种改良的最小二乘估计法，通过放弃最小二乘法的无偏性，以损失部分信息、降低精度为代价获得回归系数更为符合实际、更可靠的回归方法，对病态数据的拟合要强于最小二乘法。

算法使用

岭回归模型应用领域和线性回归类似，经常被用来做一些数值型变量的预测，类似房价预测、销售量预测、贷款额度预测、温度预测、适度预测等。

• 备注 ：该组件训练的时候 FeatureCols 和 VectorCol 是两个互斥参数，只能有一个参数来描述算法的输入特征。

  文献或出处

  [1] Hoerl, Arthur E., and Robert W. Kennard. “Ridge regression: Biased estimation for nonorthogonal problems.” Technometrics 12.1 (1970): 55-67.
  [2] https://baike.baidu.com/item/岭回归/554917?fr=aladdin

  参数说明

代码示例

Python 代码

from pyalink.alink import *
import pandas as pd
useLocalEnv(1)
df = pd.DataFrame([
    [2, 1, 1],
    [3, 2, 1],
    [4, 3, 2],
    [2, 4, 1],
    [2, 2, 1],
    [4, 3, 2],
    [1, 2, 1],
    [5, 3, 3]])
batchData = BatchOperator.fromDataframe(df, schemaStr='f0 int, f1 int, label int')
colnames = ["f0","f1"]
ridge = RidgeRegression()\
            .setFeatureCols(colnames)\
            .setLambda(0.1)\
            .setLabelCol("label")\
            .setPredictionCol("pred")
model = ridge.fit(batchData)
model.transform(batchData).print()

Java 代码

import org.apache.flink.types.Row;
import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import com.alibaba.alink.pipeline.regression.RidgeRegression;
import com.alibaba.alink.pipeline.regression.RidgeRegressionModel;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class RidgeRegressionTest {
    @Test
    public void testRidgeRegression() throws Exception {
        List <Row> df = Arrays.asList(
            Row.of(2, 1, 1),
            Row.of(3, 2, 1),
            Row.of(4, 3, 2),
            Row.of(2, 4, 1),
            Row.of(2, 2, 1),
            Row.of(4, 3, 2),
            Row.of(1, 2, 1)
        );
        BatchOperator <?> batchData = new MemSourceBatchOp(df, "f0 int, f1 int, label int");
        String[] colnames = new String[] {"f0", "f1"};
        RidgeRegression ridge = new RidgeRegression()
            .setFeatureCols(colnames)
            .setLambda(0.1)
            .setLabelCol("label")
            .setPredictionCol("pred");
        RidgeRegressionModel model = ridge.fit(batchData);
        model.transform(batchData).print();
    }
}

运行结果

| f0 | f1 | label | pred | | —- | —- | —- | —- |

| 2 | 1 | 1 | 0.8849 |

| 3 | 2 | 1 | 1.2828 |

| 4 | 3 | 2 | 1.6807 |

| 2 | 4 | 1 | 1.1334 |

| 2 | 2 | 1 | 0.9678 |

| 4 | 3 | 2 | 1.6807 |

| 1 | 2 | 1 | 0.6527 |