Class Confidence has less number of classes than expectedIssue Solution contains less number classes in the Class Confidence related list than expected.CauseIf the solution definition contains more than 300K records, the system automatically picks the latest max limit records.Example: The complete dataset had 70 classes; 25 classes were modelled using the training data (latest 300k); 25 classes are displayed in the class confidence associated list.ResolutionPossible workarounds to fix the issue Workaround 1: Increase the "Maximum number of records used in training" value (default is 300K) Try increasing the limit for max records. It is a glide property which can be updated: glide.platform_ml_api.csv_max_line as below Predictive IntelligenceConfiguration"Maximum number of records used in training" Caution: Be informed that increasing the data limit will have an impact on training times and model sizes. Training might fail as well due to memory issues.Test this in sub-prod instances thoroughly before following the same in production. Workaround 2: Adjust the filter conditions in the solution definition to have below 300K records covering all the classes.Related LinksPredictive Intelligence Common issues