Python pyspark RegressionEvaluator用法及代碼示例

本文簡要介紹 pyspark.ml.evaluation.RegressionEvaluator 的用法。

用法: class pyspark.ml.evaluation.RegressionEvaluator(*, predictionCol='prediction', labelCol='label', metricName='rmse', weightCol=None, throughOrigin=False)

回歸評估器，它需要輸入列預測、標簽和可選的權重列。

1.4.0 版中的新函數。

例子：

>>> scoreAndLabels = [(-28.98343821, -27.0), (20.21491975, 21.5),
...   (-25.98418959, -22.0), (30.69731842, 33.0), (74.69283752, 71.0)]
>>> dataset = spark.createDataFrame(scoreAndLabels, ["raw", "label"])
...
>>> evaluator = RegressionEvaluator()
>>> evaluator.setPredictionCol("raw")
RegressionEvaluator...
>>> evaluator.evaluate(dataset)
2.842...
>>> evaluator.evaluate(dataset, {evaluator.metricName: "r2"})
0.993...
>>> evaluator.evaluate(dataset, {evaluator.metricName: "mae"})
2.649...
>>> re_path = temp_path + "/re"
>>> evaluator.save(re_path)
>>> evaluator2 = RegressionEvaluator.load(re_path)
>>> str(evaluator2.getPredictionCol())
'raw'
>>> scoreAndLabelsAndWeight = [(-28.98343821, -27.0, 1.0), (20.21491975, 21.5, 0.8),
...   (-25.98418959, -22.0, 1.0), (30.69731842, 33.0, 0.6), (74.69283752, 71.0, 0.2)]
>>> dataset = spark.createDataFrame(scoreAndLabelsAndWeight, ["raw", "label", "weight"])
...
>>> evaluator = RegressionEvaluator(predictionCol="raw", weightCol="weight")
>>> evaluator.evaluate(dataset)
2.740...
>>> evaluator.getThroughOrigin()
False

相關用法

注：本文由純淨天空篩選整理自spark.apache.org大神的英文原創作品 pyspark.ml.evaluation.RegressionEvaluator。非經特殊聲明，原始代碼版權歸原作者所有，本譯文未經允許或授權，請勿轉載或複製。