[SPARK-11918] Better error from WLS for cases like singular input

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Component/s: ML
• Assignee: Sean R. Owen
• Reporter: Yanbo Liang
• Created: 23/Nov/15 07:59
• Updated: 21/Sep/16 18:59
• Resolved: 21/Sep/16 18:59

Description

Weighted Least Squares (WLS) is one of the optimization method for solve Linear Regression (when #feature < 4096). But if the dataset is very ill condition (such as 0-1 based label used for classification and the equation is underdetermined), the WLS failed (But "l-bfgs" can train and get the model). The failure is caused by the underneath lapack library return error value when Cholesky decomposition.
This issue is easy to reproduce, you can train a LinearRegressionModel by "normal" solver with the example dataset(https://github.com/apache/spark/blob/master/data/mllib/sample_libsvm_data.txt). The following is the exception:

assertion failed: lapack.dpotrs returned 1.
java.lang.AssertionError: assertion failed: lapack.dpotrs returned 1.
	at scala.Predef$.assert(Predef.scala:179)
	at org.apache.spark.mllib.linalg.CholeskyDecomposition$.solve(CholeskyDecomposition.scala:42)
	at org.apache.spark.ml.optim.WeightedLeastSquares.fit(WeightedLeastSquares.scala:117)
	at org.apache.spark.ml.regression.LinearRegression.train(LinearRegression.scala:180)
	at org.apache.spark.ml.regression.LinearRegression.train(LinearRegression.scala:67)
	at org.apache.spark.ml.Predictor.fit(Predictor.scala:90)

Issue Links

• is duplicated by: SPARK-17588 java.lang.AssertionError: assertion failed: lapack.dppsv returned 105. when running glm using gaussian link function.
• links to: [Github] Pull Request #15177 (srowen)