排序评估 (EvalRankingBatchOp)

Java 类名：com.alibaba.alink.operator.batch.evaluation.EvalRankingBatchOp
Python 类名：EvalRankingBatchOp

功能介绍

排序评估是对推荐排序算法的预测结果进行效果评估，支持下列评估指标。

hitRate

#### averageReciprocalHitRank #### map (Mean Average Precision) #### ndcgArray (Normalized Discounted Cumulative Gain)

subsetAccuracy

#### hammingLoss #### accuracy #### microPrecision #### microRecall #### microF1 #### precision #### recall #### f1

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 | | —- | —- | —- | —- | —- | —- | —- |

| labelCol | 标签列名 | 输入表中的标签列名 | String | ✓ | | |

| predictionCol | 预测结果列名 | 预测结果列名 | String | ✓ | | |

| labelRankingInfo | Object列列名 | Object列列名 | String | | | “object” |

| predictionRankingInfo | Object列列名 | Object列列名 | String | | | “object” |

代码示例

Python 代码

from pyalink.alink import *
import pandas as pd
useLocalEnv(1)
df = pd.DataFrame([
    ["{\"object\":\"[1, 6, 2, 7, 8, 3, 9, 10, 4, 5]\"}", "{\"object\":\"[1, 2, 3, 4, 5]\"}"],
    ["{\"object\":\"[4, 1, 5, 6, 2, 7, 3, 8, 9, 10]\"}", "{\"object\":\"[1, 2, 3]\"}"],
    ["{\"object\":\"[1, 2, 3, 4, 5]\"}", "{\"object\":\"[]\"}"]
])
inOp = BatchOperator.fromDataframe(df, schemaStr='pred string, label string')
metrics = EvalRankingBatchOp().setPredictionCol('pred').setLabelCol('label').linkFrom(inOp).collectMetrics()
print(metrics)

Java 代码

import org.apache.flink.types.Row;
import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.evaluation.EvalRankingBatchOp;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import com.alibaba.alink.operator.common.evaluation.RankingMetrics;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class EvalRankingBatchOpTest {
    @Test
    public void testEvalRankingBatchOp() throws Exception {
        List <Row> df = Arrays.asList(
            Row.of("{\"object\":\"[1, 6, 2, 7, 8, 3, 9, 10, 4, 5]\"}", "{\"object\":\"[1, 2, 3, 4, 5]\"}"),
            Row.of("{\"object\":\"[4, 1, 5, 6, 2, 7, 3, 8, 9, 10]\"}", "{\"object\":\"[1, 2, 3]\"}"),
            Row.of("{\"object\":\"[1, 2, 3, 4, 5]\"}", "{\"object\":\"[]\"}")
        );
        BatchOperator <?> inOp = new MemSourceBatchOp(df, "pred string, label string");
        RankingMetrics metrics = new EvalRankingBatchOp().setPredictionCol("pred").setLabelCol("label").linkFrom(inOp)
            .collectMetrics();
        System.out.println(metrics.toString());
    }
}

运行结果

-------------------------------- Metrics: --------------------------------
microPrecision:0.32
averageReciprocalHitRank:0.5
precision:0.2667
accuracy:0.2667
f1:0.3761
hitRate:0.6667
microRecall:1
microF1:0.4848
subsetAccuracy:0
recall:0.6667
map:0.355
hammingLoss:0.5667