Error Reported When Parameter Values of a Machine Learning Algorithm Are Parsed

Symptom

When a machine learning algorithm library is running, the log shows that the parameters prefixed with spark.boostkit are incorrectly parsed. Then, the program automatically stops. For example, if the parameters of the PrefixSpan algorithm are incorrectly set, the following log information will be displayed:

parse boostkit parameter(spark.boostkit.ml.ps.localTimeout) failed, Error reason: ......

Key Process and Cause Analysis

The spark conf parameters newly added to the algorithms are all prefixed with spark.boostkit. If the parameter values do not meet the requirements, a parsing error occurs.

Conclusion and Solution

Set the parameter values correctly. For details, see Algorithm APIs in the Machine Learning Algorithm Library Developer Guide.

Parent topic: Troubleshooting