Java 类名:com.alibaba.alink.operator.batch.outlier.IForestOutlier4GroupedDataBatchOp
Python 类名:IForestOutlier4GroupedDataBatchOp
iForest 可以识别数据中异常点,在异常检测领域有比较好的效果。算法使用 sub-sampling 方法,降低了算法的计算复杂度。
| 名称 | 中文名称 | 描述 | 类型 | 是否必须? | 取值范围 | 默认值 |
|---|---|---|---|---|---|---|
| inputMTableCol | 输入列名 | 输入序列的列名 | String | ✓ | ||
| outputMTableCol | 输出列名 | 输出序列的列名 | String | ✓ | ||
| predictionCol | 预测结果列名 | 预测结果列名 | String | ✓ | ||
| featureCols | 特征列名数组 | 特征列名数组,默认全选 | String[] | 所选列类型为 [BIGDECIMAL, BIGINTEGER, BYTE, DOUBLE, FLOAT, INTEGER, LONG, SHORT] | null | |
| maxOutlierNumPerGroup | 每组最大异常点数目 | 每组最大异常点数目 | Integer | |||
| maxOutlierRatio | 最大异常点比例 | 算法检测异常点的最大比例 | Double | |||
| numTrees | 模型中树的棵数 | 模型中树的棵数 | Integer | 100 | ||
| outlierThreshold | 异常评分阈值 | 只有评分大于该阈值才会被认为是异常点 | Double | |||
| predictionDetailCol | 预测详细信息列名 | 预测详细信息列名 | String | |||
| subsamplingSize | 每棵树的样本采样行数 | 每棵树的样本采样行数,默认 256 ,最小 2 ,最大 100000 . | Integer | 1 <= x <= 100000 | 256 | |
| tensorCol | tensor列 | tensor列 | String | 所选列类型为 [BOOL_TENSOR, BYTE_TENSOR, DOUBLE_TENSOR, FLOAT_TENSOR, INT_TENSOR, LONG_TENSOR, STRING, STRING_TENSOR, TENSOR, UBYTE_TENSOR] | null | |
| vectorCol | 向量列名 | 向量列对应的列名,默认值是null | String | 所选列类型为 [DENSE_VECTOR, SPARSE_VECTOR, STRING, VECTOR] | null | |
| numThreads | 组件多线程线程个数 | 组件多线程线程个数 | Integer | 1 |
import pandas as pd
df = pd.DataFrame([
[1, 1, 10.0],
[1, 2, 11.0],
[1, 3, 12.0],
[1, 4, 13.0],
[1, 5, 14.0],
[1, 6, 15.0],
[1, 7, 16.0],
[1, 8, 17.0],
[1, 9, 18.0],
[1, 10, 19.0]
])
dataOp = BatchOperator.fromDataframe(
df, schemaStr='group_id int, id int, val double')
outlierOp = dataOp.link(
GroupByBatchOp()
.setGroupByPredicate("group_id")
.setSelectClause("mtable_agg(id, val) as data")
).link(
IForestOutlier4GroupedDataBatchOp()
.setInputMTableCol("data")
.setOutputMTableCol("pred")
.setFeatureCols(["val"])
.setPredictionCol("detect_pred")
)
outlierOp.print()
import org.apache.flink.types.Row;
import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.outlier.IForestOutlier4GroupedDataBatchOp;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import com.alibaba.alink.operator.batch.sql.GroupByBatchOp;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class IForestOutlier4GroupedDataBatchOpTest {
@Test
public void test() throws Exception {
List <Row> mTableData = Arrays.asList(
Row.of(1, 1, 10.0),
Row.of(1, 2, 11.0),
Row.of(1, 3, 12.0),
Row.of(1, 4, 13.0),
Row.of(1, 5, 14.0),
Row.of(1, 6, 15.0),
Row.of(1, 7, 16.0),
Row.of(1, 8, 17.0),
Row.of(1, 9, 18.0),
Row.of(1, 10, 19.0)
);
MemSourceBatchOp dataOp = new MemSourceBatchOp(mTableData, new String[] {"group_id", "id", "val"});
BatchOperator <?> outlierOp = dataOp.link(
new GroupByBatchOp()
.setGroupByPredicate("group_id")
.setSelectClause("group_id, mtable_agg(id, val) as data")
).link(
new IForestOutlier4GroupedDataBatchOp()
.setInputMTableCol("data")
.setOutputMTableCol("pred")
.setFeatureCols("val")
.setPredictionCol("detect_pred")
);
outlierOp.print();
}
}
| group_id | data | pred | |||
|---|---|---|---|---|---|
| 1 | MTable(10,2)(id,val) | MTable(10,3)(id,val,detect_pred) | |||
| 1 | 10.0000 | 1 | 10.0000 | false | |
| 2 | 11.0000 | 2 | 11.0000 | false | |
| 3 | 12.0000 | 3 | 12.0000 | false | |
| 4 | 13.0000 | 4 | 13.0000 | false | |
| 5 | 14.0000 | 5 | 14.0000 | false |