Icon

Heart Disease - Machine Learning Case - Comparing Algorithms

Binary Classification - use Python XGBoost package and other nodes to build model and deploy that thru KNIME Python nodes<br /><br />prepare data with vtreat package<br />in the subfolder /data/notebooks/ there is a Jupyter notebook to experiment and build XGBoost models ("kn_example_python_xgboost.ipynb")<br /><br />Also you can further explore the H2O.ai AutoML model with the notebook "h2o_inspect_model_automl_existing.ipynb"<br /><br />------------<br />Heart Failure Prediction Dataset (Kaggle)<br />11 clinical features for predicting heart disease events.<br /><br />https://www.kaggle.com/datasets/fedesoriano/heart-failure-prediction?resource=download

URL: How to Develop Your First XGBoost Model in Python https://machinelearningmastery.com/develop-first-xgboost-model-python-scikit-learn/
URL: A Beginner’s guide to XGBoost https://towardsdatascience.com/a-beginners-guide-to-xgboost-87f5d4c30ed7
URL: XGBoost Parameters https://xgboost.readthedocs.io/en/stable/parameter.html
URL: forum entry (45057) https://forum.knime.com/t/saving-xgboost-model-to-pmml-possible-now/45057/4?u=mlauber71
URL: Meta Collection about KNIME and Python https://kni.me/w/AvjrddXKOIoZYLV3
URL: Medium: Data preparation for Machine Learning with KNIME and the Python “vtreat” package https://medium.com/p/efcaf58fa783
URL: H2O.ai AutoML (wrapped with Python) in KNIME for classification problems https://forum.knime.com/t/h2o-ai-automl-in-knime-for-classification-problems/20923?u=mlauber71
URL: HUB: Binary Classification - Heart Disease - Machine Learning Case - Comparing Algorithms https://hub.knime.com/-/spaces/-/~0dXvsD0vMrv_w6Fw/current-state/
URL: forum entry (77228) https://forum.knime.com/t/how-do-know-if-your-dataset-is-good-before-using-ml-algorithms/77228/7?u=mlauber71
URL: Medium Blog: KNIME — Machine Learning and Artificial Intelligence— A Collection https://medium.com/p/12e0f7d83b50
URL: Medium Blog: About Machine-Learning — How it Fails and Succeeds https://medium.com/p/9f3ab7cb9b00
URL: Medium Blog: KNIME, XGBoost and Optuna for Hyper Parameter Optimization https://medium.com/p/dcf0efdc8ddf

Binary Classification - use Python XGBoost package and other nodes to build model and deploy that thru KNIME Python nodes

also: prepare data with vtreat package
in the subfolder /data/ there is a Jupyter notebook to experiment and build XGBoost models ("kn_example_python_xgboost.ipynb"). Also a Notebook to extract H2O.ai parameters from a model ("h2o_inspect_model_automl_existing.ipynb")

This workflow has been adapted for KNIME 5.2+ from the original one for 4.7 (https://hub.knime.com/-/spaces/-/~In0Rxt7EhzQfycx3/current-state/)

in the subfolder /data/ there is a Jupyter notebook to experiment and build XGBoost models ("kn_example_python_xgboost.ipynb")

This is what deployment would look like on new data using the model from inside Python

Python Conda environment propagation. Please read this article for more details:


KNIME and Python — Setting up and managing Conda environments
https://medium.com/p/2ac217792539

Medium: Data preparation for Machine Learning with KNIME and the Python “vtreat” package

https://medium.com/p/efcaf58fa783
https://forum.knime.com/t/data-preparation-for-machine-learning-with-knime-and-the-python-vtreat-package/58679?u=mlauber71

About Machine-Learning — How it Fails and Succeeds

https://medium.com/p/9f3ab7cb9b00

  

KNIME — Machine Learning and Artificial Intelligence— A Collection

https://medium.com/p/12e0f7d83b50

MEDIUM Blog: Hyperparameter optimization for LightGBM — wrapped in KNIME nodes

https://medium.com/p/ddb7ae1d7e2


GitHub Repository

https://github.com/ml-score/knime_meets_python/tree/main/machine_learning/binary

train.parquetTarget 0/1 - target variablerow_id = ID column
Parquet Reader
sort by AUC (pr)DESCENDING
Sorter
collect the results
Concatenate
$$ROWID$$ = "eclipse.home.location" => TRUE
Rule-based Row Filter
Table Row to Variable
inspect the model
Binary Classification Inspector
Py_XGBoost
Prepare Heart Failure Prediction Dataset (Kaggle)https://www.kaggle.com/datasets/fedesoriano/heart-failure-prediction/
Component
knime_model_gbm
also variable importance
h2o_glm
also variable importance
knime_xgboost_model
also variable importance
h2o_glm_vtreat
$$ROWID$$ = "java.home" => TRUE
Rule-based Row Filter
drop Targetas if this werecompletely new data
Column Filter
without DL
h2o_automl
SVM usingthe normalized data
knime_libsvm
Apply the XGBoost model from"knime_xgboost_model.json"
Python Script
without DL
h2o_automl_vtreat
SVM usingthe normalized data
knime_svm
knime_model_gbm_vtreat
test_normalized.table
Table Writer
knime_xgboost_model_vtreat
train_normalized.table
Table Writer
v_model_json_file"knime_xgboost_model.json"
Java Edit Variable (simple)
osgi.syspath=> find where theKNIME installtion is locatedon Windows or MacOSX also java.home
Extract System Properties
model_results.xlsx
Excel Writer
RowID
also preparingdata with normalization
BINARY_RPROP_MLP
locate and create /data/ folder with absolute paths
Collect Local Metadata
export Flow Variables from KNIME ^(?!knime.workspace).*$
Variable to Table Row
Py_XGBoost_vtreat
determine package versions
Python Script
Merge Variables
v_rounds_to_run=> edit!return 2500;
Java Edit Variable (simple)
Table Transposer
also variable importance
knime_ranfor
knime_ranfor_vtreat
just Deep Learning
h2o_automl_dl_vtreat
Binary Classification Inspector
just Deep Learning
h2o_automl_dl
Medium: Data preparation for Machine Learning withKNIME and the Python “vtreat” packagehttps://medium.com/p/efcaf58fa783
vtreat prepare binary data
Activate Conda Environmentbased on Operating SystemWindows or macOS
conda_environment_kaggle
jupyter_test_prediction.parquet
Parquet Reader
test.parquetTarget 0/1 - target variablerow_id = ID column
Parquet Reader
v_vtreat_indicator_min_fraction=> edit!return 0.025;https://github.com/WinVector/pyvtreat/blob/main/Examples/Classification/Classification.md
Java Edit Variable (simple)
edit: v_runtime_automlset the maximum runtime ofH2O.ai AutoML in SECONDS600
Integer Input (legacy)

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: m_lauber
Created at: 2022-07-13
On NodePit since: 2024-02-16
Last update: 2026-03-01
Created with KNIME version: v5.10.0
Tags: xgboostpythonmodelknimejupytervtreatsvmh2o.aih2obinarygbm

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

Try NodePit Runner!