Question

是否有可能在Apache Spark中的多类分类问题中找到错误度量（精度和召回率）。我正在使用来自Spark的MlLib的Logistic回归来构建我的模型，并想使用错误度量来评估我的模型。

Answer 1

假设您的测试数据在$.ajax({ url: queryURL, method: "GET", }).done(function(response) { var events = response.events; for (var i = 0; i < 3; i++) { var tRow = $("<tr>"); var url = events[i].url console.log(events[i].url); tRow.append($("<td>").text(events[i].name.text)); tRow.append($("<td>").html(url)); tRow.append($("<td>").text(events[i].start.local)); tRow.append($("<td>").text(events[i].end.local)); $("#eventTable").append(tRow); } eventDB.push({ event, address, distance, })

中

test

混乱矩阵

import org.apache.spark.mllib.classification.LogisticRegressionWithLBFGS
import org.apache.spark.mllib.evaluation.MulticlassMetrics
import org.apache.spark.mllib.regression.LabeledPoint
import org.apache.spark.mllib.util.MLUtils

val predictionAndLabels = test.map { case LabeledPoint(label,
   features) =>
     val prediction = model.predict(features)
     (prediction, label)
}

val metrics = new MulticlassMetrics(predictionAndLabels)

总体统计

println("Confusion matrix:")
println(metrics.confusionMatrix)

按标签精度

val accuracy = metrics.accuracy
println("Summary Statistics")
println(s"Accuracy = $accuracy")

按标签调用

val labels = metrics.labels
labels.foreach { l =>
  println(s"Precision($l) = " + metrics.precision(l))
}

标签误报率

labels.foreach { l =>
  println(s"Recall($l) = " + metrics.recall(l))
}

按标签的F量测

labels.foreach { l =>
  println(s"FPR($l) = " + metrics.falsePositiveRate(l))
}

加权统计

labels.foreach { l =>
  println(s"F1-Score($l) = " + metrics.fMeasure(l))
}

Apache Spark中的多类分类错误指标

1 个答案: