Icon

COS10022_​A1_​Workflow_​Darren Pui Wei Lik

Step 1-10

Step 11-14

Step 15-18

Step 19-20

Model D: Predict on Suite 2 test set
Logistic Regression Predictor
Model L: Collect predictions from all 10 folds
X-Aggregator
Model L: Evaluate cross-validated performance
Scorer
Model F: Predict on Suite 2 test set
Random Forest Predictor
Model M: Predict on Leave-one-out fold
Logistic Regression Predictor
Model D: Evaluate precision, recall, F1, accuracy
Scorer
Model M: Collect predictions from Leave-one-out folds
X-Aggregator
Step 17: Stratified Leave-one-out cross-validation, seed 9698
X-Partitioner
Model M: Logistic Regression, ref class 1, SAG solver, 10000 epochs, seed 9698
Logistic Regression Learner
Model E: Predict on Suite 2 test set
Naive Bayes Predictor
Step 19: Training-set performance for overfitting check
Scorer
Model H: Naive Bayes, Suite 3 (90/10), default prob 0.0001, min std 0.0001
Naive Bayes Learner
Step 19: ROC curve on training set
ROC Curve
Model F: Evaluate precision, recall, F1, accuracy
Scorer
Model M: Evaluate cross-validated performance
Scorer
Model E: Naive Bayes, Suite 2 (70/30), default prob 0.0001, min std 0.0001
Naive Bayes Learner
Step 19: Predict on TRAINING set using the same best model from Step 11
Logistic Regression Predictor
Model H: Evaluate precision, recall, F1, accuracy
Scorer
Model I: Random Forest, Suite 3 (90/10), Gini, 200 trees, seed 9698
Random Forest Learner
Model H: Predict on Suite 3 test set
Naive Bayes Predictor
Model G: Logistic Regression, Suite 3 (90/10), ref class 1, SAG solver, 10000 epochs, seed 9698
Logistic Regression Learner
set 1.0E-5 for Logistic Regression Learner
Variable Creator
Step 19: ROC curve on test set
ROC Curve
Model I: Predict on Suite 3 test set
Random Forest Predictor
Model G: Evaluate precision, recall, F1, accuracy
Scorer
Model G: Predict on Suite 3 test set
Logistic Regression Predictor
Step2: Convert 5 binary attributes and the target DEATH_EVENT from Number to String
Number to String
Step3: Z-score normalization of 7 numerical attributes
Normalizer
Model I: Evaluate precision, recall, F1, accuracy
Scorer
Step1：Load heart failure dataset
CSV Reader
Model J: Predict on Suite 1 test set, random sampling
Logistic Regression Predictor
Step 12: Random split using best ratio from Step 11, seed 9698
Table Partitioner
Model J: Logistic Regression, Suite 1 (50/50), ref class 1, SAG solver, 10000 epochs, seed 9698, random sampling
Logistic Regression Learner
Step1：Load heart failure dataset
CSV Reader
Step2: Convert 5 binary attributes and the target DEATH_EVENT from Number to String
Number to String
Step3: Z-score normalization of 7 numerical attributes
Normalizer
Step 4: Suite 1 - 50% train
Table Partitioner
Step4 :Suite 2 - 70% train
Table Partitioner
Step4: Suite 3 - 90% train
Table Partitioner
Model A: Logistic Regression, Suite 1 (50/50), ref class 1, SAG solver, 10000 epochs, seed 9698
Logistic Regression Learner
Step 13: Random feature subsets 2-12 features, 50 iterations, seed 9698
Feature Selection Loop Start (1:1)
Model J: Evaluate precision, recall, F1, accuracy, random sampling
Scorer
Predict on inner test set
Random Forest Predictor
Inner accuracy for each feature subset
Scorer
Stratified 80/20 split inside the loop, seed 9698
Table Partitioner
Random Forest replaces Naive Bayes: Gini, 200 trees, seed 9698
Random Forest Learner
Collect accuracy of each feature subset
Feature Selection Loop End
Keep the feature subset with the best accuracy
Feature Selection Filter
Step 13: Keep only the features selected by Forward Feature Selection
Reference Column Filter
Model K: Predict on Suite 1 test set on selected features
Logistic Regression Predictor
Model B: Naive Bayes, Suite 1 (50/50), default prob 0.0001, min std 0.0001
Naive Bayes Learner
Model K: Evaluate precision, recall, F1, accuracy on selected features
Scorer
Model K: stratified split using best ratio, seed 9698
Table Partitioner
Model K: Logistic Regression, Suite 1 (50/50), ref class 1, SAG solver, 10000 epochs, seed 9698 on selected features
Logistic Regression Learner
Model A: Predict on Suite 1 test set
Logistic Regression Predictor
Model C: Random Forest, Suite 1 (50/50), Gini, 200 trees, seed 9698
Random Forest Learner
Step 16: Stratified 10-fold cross-validation, seed 9698
X-Partitioner
Model A: Evaluate precision, recall, F1, accuracy
Scorer
Model C: Predict on Suite 1 test set
Random Forest Predictor
Model L: Logistic Regression, ref class 1, SAG solver, 10000 epochs, seed 9698 trained on 9 folds
Logistic Regression Learner
Model B: Predict on Suite 1 test set
Naive Bayes Predictor
Step 14: ROC curve for Model J (random sampling)
ROC Curve
Model B: Evaluate precision, recall, F1, accuracy
Scorer
Step 14: ROC curve for Model K (selected features)
ROC Curve
Model E: Evaluate precision, recall, F1, accuracy
Scorer
Model F: Random Forest, Suite 2 (70/30), Gini, 200 trees, seed 9698
Random Forest Learner
Model C: Evaluate precision, recall, F1, accuracy
Scorer
Model L: Predict on held-out fold
Logistic Regression Predictor
Model D: Logistic Regression, Suite 2 (70/30), ref class 1, SAG solver, 10000 epochs, seed 9698
Logistic Regression Learner

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: User
Created at: 2026-10-13
On NodePit since: 2026-10-18
Last update: 2026-10-18
Created with KNIME version: v5.12.0

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

Try NodePit Runner!