Can someone point me to the sparklyr equivalent of predict.lm's prediction and confidence interval arguments?
With predict.lm for example, one can use:
predict(fit, newdata = test_data, interval = "pred", level = 0.9)
And you get a prediction interval for every record, but I'm not seeing anything similar for ml_pipelines or sdf_predict or ml_fit or ml_transform. So, my pipelines get the point estimate, but I have no way of getting the intervals.