Icon

Exercise2

<p><strong>Chapter4/Exercise 2</strong></p><p>This is an exercise for model improvement by changing the classification thresholds. In the previous exercise "Chapter 4/Exercise 1" some patterns were not correctly classified. By applying a different set of thresholds to the output probabilities, the final wine classification can be improved.</p><p>The key node of this exercise is the <em>Expression </em>node at the end, defining a set of thresholds for a more exact classification.</p>

URL: KNIME Beginner's Luck (Book Homepage) https://www.knime.com/knimepress/beginners-luck

Workflow: Chapter 4/Exercise 2

This is an exercise for model improvement by changing the classification thresholds. In the previous exercise "Chapter 4/Exercise 1" some patterns were not correctly classified. By applying a different set of thresholds to the output probabilities, the final wine classification can be improved.

The key node of this exercise is the Expression node at the end, defining a set of thresholds for a more exact classification.

Reading data

Transforming data

Training model

Applying trained model

Evaluating model

Postprocessing data

80% vs. 20%
Table Partitioner
Confusion matrixClass vs. Prediction
Scorer
Convert "Class"to String
Expression
14-10-3
RProp MLP Learner
More conservative classes:if pred class > 0.7 => class 3,if pred class < 0.3 => class 1,otherwise class 2
Expression
MultiLayerPerceptron Predictor
ApplyMin-Max-Normalization
Normalizer (Apply)
Min-Max-Normalizationin [0,1]
Normalizer
wine.csv
CSV Reader
Class back to 1,2,3
Denormalizer

Nodes

Extensions

Links

Download

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

Download Workflow
On NodePit since: 2025-12-15
Last update: 2026-01-05
Created with KNIME version: v5.9.0
Tags: EducationKNIMEPressMLPPartitionNormalizationScorer

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

Try NodePit Runner!