Icon

COPA 2023 - Mondrian conformal predictive systems experiment

This workflow has been prepared as part of a paper introducing conformal predictive systems in KNIME. The paper was submitted to the conference COPA 2023.


Mondrian CPS withoutnormalization. Normalized Mondrian CPS usingthe variance from the Randomforest as Sigma. CPS without normalization. Normalized CPS using thevariance from the Random forestas Sigma. Calibration10-fold CVTestLearn targetLoad the path to datasets stored in the projects Data folderDatasetsTargetCalibration=1/3Training=2/3Pathend repetitionsRepetitionsend CVDatasetPredictionPredictionMondrianPrediction [Binned]Prediction [Binned]end MondrianLoaddatasetDatasetend DatasetNormalizedStandardDatasetAverage RegressionPerformanceRegressionPerformanceper datasetRemove too small datasetsBinned Median vs targetError rates and interval widthsper datasetNormalized MondrianMondrianBinned Median vs targetError rates and interval widthsNode 564Node 565Node 567 Random Forest Predictor(Regression) X-Partitioner Random Forest Predictor(Regression) Random Forest Learner(Regression) List Files/Folders Table Row ToVariable Loop Start Normalizer Path to String ConformalPartitioning String Manipulation(Variable) Loop End Counting Loop Start Loop End CSV Writer ConstantValue Column Auto-Binner Auto-Binner (Apply) Group Loop Start ReferenceRow Filter Loop End Scoring Setup the intervalpercentiles File Reader ConstantValue Column Loop End Predictive SystemsRegression Predictive SystemsRegression Numeric Scoring ConstantValue Column GroupBy Pivoting Rule-basedRow Filter Line Plot Error rate visualization- Regression Pivoting Predictive SystemsRegression Predictive SystemsRegression Bar Chart Result Cleanup Result Cleanup Pivoting Column Resorter Column Resorter Result Cleanup Aggregating onPrediction [Binned] Mondrian CPS withoutnormalization. Normalized Mondrian CPS usingthe variance from the Randomforest as Sigma. CPS without normalization. Normalized CPS using thevariance from the Random forestas Sigma. Calibration10-fold CVTestLearn targetLoad the path to datasets stored in the projects Data folderDatasetsTargetCalibration=1/3Training=2/3Pathend repetitionsRepetitionsend CVDatasetPredictionPredictionMondrianPrediction [Binned]Prediction [Binned]end MondrianLoaddatasetDatasetend DatasetNormalizedStandardDatasetAverage RegressionPerformanceRegressionPerformanceper datasetRemove too small datasetsBinned Median vs targetError rates and interval widthsper datasetNormalized MondrianMondrianBinned Median vs targetError rates and interval widthsNode 564Node 565Node 567Random Forest Predictor(Regression) X-Partitioner Random Forest Predictor(Regression) Random Forest Learner(Regression) List Files/Folders Table Row ToVariable Loop Start Normalizer Path to String ConformalPartitioning String Manipulation(Variable) Loop End Counting Loop Start Loop End CSV Writer ConstantValue Column Auto-Binner Auto-Binner (Apply) Group Loop Start ReferenceRow Filter Loop End Scoring Setup the intervalpercentiles File Reader ConstantValue Column Loop End Predictive SystemsRegression Predictive SystemsRegression Numeric Scoring ConstantValue Column GroupBy Pivoting Rule-basedRow Filter Line Plot Error rate visualization- Regression Pivoting Predictive SystemsRegression Predictive SystemsRegression Bar Chart Result Cleanup Result Cleanup Pivoting Column Resorter Column Resorter Result Cleanup Aggregating onPrediction [Binned]

Nodes

Extensions

Links

Download

To use this workflow in KNIME, download it from the below URL and open it in KNIME:

Download Workflow
Created by: loftuw
Created at: 2021-11-30
On NodePit since: 2023-04-08
Last update: 2026-01-17
Created with KNIME version: v4.7.2
Tags: MondrianConformal predictive systemsCOPARegressionUncertainty quantificationConformal prediction

Deploy, schedule, execute, and monitor your KNIME workflows locally, in the cloud or on-premises – with our brand new NodePit Runner.

Try NodePit Runner!