在 Apache Spark Logistic 回归算法中获得相同的精度、召回率和 F 分数值

Question

我已经针对分类问题实施了逻辑回归。我得到相同的准确率、召回率和 F1 分数值。具有相同的值可以吗？我在实施决策树和随机森林时也遇到了这个问题。我也得到了相同的精度值、召回率和 F1 分数。

// Run training algorithm to build the model.
        final LogisticRegressionModel model = new LogisticRegressionWithLBFGS()
                .setNumClasses(13).
                run(data.rdd());
//Compute raw scores on the test set.
        JavaRDD<Tuple2<Object, Object>> predictionAndLabels = testData.map(
                new Function<LabeledPoint, Tuple2<Object, Object>>() {
                    public Tuple2<Object, Object> call(LabeledPoint p) {
                        Double prediction = model.predict(p.features());
                        return new Tuple2<Object, Object>(prediction, p.label());
                    }
                }
        );
// Get evaluation metrics.
        MulticlassMetrics metrics = new MulticlassMetrics(predictionAndLabels.rdd());
        double precision = metrics.precision();
        System.out.println("Precision = " + precision);

        double recall = metrics.recall();
        System.out.println("Recall = " + recall);

        double FScore = metrics.fMeasure();
        System.out.println("F Measure = " + FScore);

Answer 1

我也遇到了同样的问题。我尝试过决策树、随机森林和 GBT。每次，我都得到相同的精度、召回率和 F1 分数。准确率也一样（通过混淆矩阵计算）

所以，我使用自己的公式和编写的代码来获得准确度、精确度、召回率和 F1 分数度量。

from pyspark.ml.classification import RandomForestClassifier
from pyspark.mllib.evaluation import MulticlassMetrics

#generate model on splited dataset
rf = RandomForestClassifier(labelCol='label', featuresCol='features')
fit = rf.fit(trainingData)
transformed = fit.transform(testData)

results = transformed.select(['prediction', 'label'])
predictionAndLabels=results.rdd
metrics = MulticlassMetrics(predictionAndLabels)

cm=metrics.confusionMatrix().toArray()
accuracy=(cm[0][0]+cm[1][1])/cm.sum()
precision=(cm[0][0])/(cm[0][0]+cm[1][0])
recall=(cm[0][0])/(cm[0][0]+cm[0][1])`
print("RandomForestClassifier: accuracy,precision,recall",accuracy,precision,recall)

Answer 2

您可以将 label=1 作为二进制 classification 的精度和召回方法的参数。它对我有用。对于多个 classification，您可以尝试计算精度和召回值的 class 的标签索引。

`double precision = metrics.precision(label=1);
 System.out.println("Precision = " + precision);
 double recall = metrics.recall(label=1);
 System.out.println("Recall = " + recall);
 double FScore = metrics.fMeasure();
 System.out.println("F Measure = " + FScore);`

在 Apache Spark Logistic 回归算法中获得相同的精度、召回率和 F 分数值

Get same value for precision, recall and F score in Apache Spark Logistic regression algorithm

apache-spark

performance-measuring