Java 类名:com.alibaba.alink.operator.batch.classification.C45TrainBatchOp
Python 类名:C45TrainBatchOp
c45是一种常用的树模型
c45组件支持稠密数据格式
支持带样本权重的训练
名称 | 中文名称 | 描述 | 类型 | 是否必须? | 取值范围 | 默认值 |
---|---|---|---|---|---|---|
labelCol | 标签列名 | 输入表中的标签列名 | String | ✓ | ||
categoricalCols | 离散特征列名 | 离散特征列名 | String[] | 所选列类型为 [BOOLEAN, DATE, DOUBLE, FLOAT, INTEGER, LONG, SHORT, STRING, TIME, TIMESTAMP] | ||
createTreeMode | 创建树的模式。 | series表示每个单机创建单颗树,parallel表示并行创建单颗树。 | String | “series” | ||
featureCols | 特征列名数组 | 特征列名数组,默认全选 | String[] | 所选列类型为 [BOOLEAN, DATE, DOUBLE, FLOAT, INTEGER, LONG, SHORT, STRING, TIME, TIMESTAMP] | null | |
maxBins | 连续特征进行分箱的最大个数 | 连续特征进行分箱的最大个数。 | Integer | 128 | ||
maxDepth | 树的深度限制 | 树的深度限制 | Integer | 2147483647 | ||
maxLeaves | 叶节点的最多个数 | 叶节点的最多个数 | Integer | 2147483647 | ||
maxMemoryInMB | 树模型中用来加和统计量的最大内存使用数 | 树模型中用来加和统计量的最大内存使用数 | Integer | 64 | ||
minInfoGain | 分裂的最小增益 | 分裂的最小增益 | Double | 0.0 | ||
minSampleRatioPerChild | 子节点占父节点的最小样本比例 | 子节点占父节点的最小样本比例 | Double | 0.0 | ||
minSamplesPerLeaf | 叶节点的最小样本个数 | 叶节点的最小样本个数 | Integer | 2 | ||
weightCol | 权重列名 | 权重列对应的列名 | String | 所选列类型为 [BIGDECIMAL, BIGINTEGER, BYTE, DOUBLE, FLOAT, INTEGER, LONG, SHORT] | null |
from pyalink.alink import * import pandas as pd useLocalEnv(1) df = pd.DataFrame([ [1.0, "A", 0, 0, 0], [2.0, "B", 1, 1, 0], [3.0, "C", 2, 2, 1], [4.0, "D", 3, 3, 1] ]) batchSource = BatchOperator.fromDataframe( df, schemaStr=' f0 double, f1 string, f2 int, f3 int, label int') streamSource = StreamOperator.fromDataframe( df, schemaStr=' f0 double, f1 string, f2 int, f3 int, label int') trainOp = C45TrainBatchOp()\ .setLabelCol('label')\ .setFeatureCols(['f0', 'f1', 'f2', 'f3'])\ .linkFrom(batchSource) predictBatchOp = C45PredictBatchOp()\ .setPredictionDetailCol('pred_detail')\ .setPredictionCol('pred') predictStreamOp = C45PredictStreamOp(trainOp)\ .setPredictionDetailCol('pred_detail')\ .setPredictionCol('pred') predictBatchOp.linkFrom(trainOp, batchSource).print() predictStreamOp.linkFrom(streamSource).print() StreamOperator.execute()
import org.apache.flink.types.Row; import com.alibaba.alink.operator.batch.BatchOperator; import com.alibaba.alink.operator.batch.classification.C45PredictBatchOp; import com.alibaba.alink.operator.batch.classification.C45TrainBatchOp; import com.alibaba.alink.operator.batch.source.MemSourceBatchOp; import com.alibaba.alink.operator.stream.StreamOperator; import com.alibaba.alink.operator.stream.classification.C45PredictStreamOp; import com.alibaba.alink.operator.stream.source.MemSourceStreamOp; import org.junit.Test; import java.util.Arrays; import java.util.List; public class C45TrainBatchOpTest { @Test public void testC45TrainBatchOp() throws Exception { List <Row> df = Arrays.asList( Row.of(1.0, "A", 0, 0, 0), Row.of(2.0, "B", 1, 1, 0), Row.of(3.0, "C", 2, 2, 1), Row.of(4.0, "D", 3, 3, 1) ); BatchOperator <?> batchSource = new MemSourceBatchOp( df, " f0 double, f1 string, f2 int, f3 int, label int"); StreamOperator <?> streamSource = new MemSourceStreamOp( df, " f0 double, f1 string, f2 int, f3 int, label int"); BatchOperator <?> trainOp = new C45TrainBatchOp() .setLabelCol("label") .setFeatureCols("f0", "f1", "f2", "f3") .linkFrom(batchSource); BatchOperator <?> predictBatchOp = new C45PredictBatchOp() .setPredictionDetailCol("pred_detail") .setPredictionCol("pred"); StreamOperator <?> predictStreamOp = new C45PredictStreamOp(trainOp) .setPredictionDetailCol("pred_detail") .setPredictionCol("pred"); predictBatchOp.linkFrom(trainOp, batchSource).print(); predictStreamOp.linkFrom(streamSource).print(); StreamOperator.execute(); } }
f0 | f1 | f2 | f3 | label | pred | pred_detail |
---|---|---|---|---|---|---|
1.0000 | A | 0 | 0 | 0 | 0 | {“0”:1.0,“1”:0.0} |
2.0000 | B | 1 | 1 | 0 | 0 | {“0”:1.0,“1”:0.0} |
3.0000 | C | 2 | 2 | 1 | 1 | {“0”:0.0,“1”:1.0} |
4.0000 | D | 3 | 3 | 1 | 1 | {“0”:0.0,“1”:1.0} |