Question

A data scientist trains a Spark ML linear regression model to predict house prices. They want to evaluate the fitted model's predictions on a held-out DataFrame using root mean squared error. They write the following code:

```python
from pyspark.ml.evaluation import RegressionEvaluator

evaluator = RegressionEvaluator(
    labelCol="price",
    predictionCol="prediction"
)
result = evaluator.evaluate(predictions_df)
```

Without specifying anything further, what value does `result` contain?

Accepted Answer

The root mean squared error (RMSE), because it is the default metric for RegressionEvaluator.

Answer

The R-squared value, because it is the default metric for RegressionEvaluator.

Answer

The mean absolute error (MAE), because it is the default metric for RegressionEvaluator.

Answer

An error is raised because the metricName parameter must always be set explicitly.