Published October 21, 2026 | Version v1

Python Code and Sample dataset for Machine Learning-Driven Model Development for Prediction of Pulmonary Tuberculosis: A Cross-Sectional Study from Ethiopia's Mid-Rift Valley

Description

This Zenodo record contains the custom Python code and a representative sample of the dataset used in the study titled "Machine Learning-Driven Model Development for Prediction of Pulmonary Tuberculosis: A Cross-Sectional Study from Ethiopia’s Mid-Rift Valley."

Technical info

  • Machine_Learning_Final_Analysis.ipynb: The complete Python code (Jupyter Notebook format) for data preprocessing, exploratory data analysis (EDA), model training and hyperparameter tuning for eight machine learning algorithms, decision threshold optimization based on Youden's J statistic, and model interpretation using SHAP and permutation importance.

  • Sample_Dataset.csv: A sample of the dataset used for the analysis. This is provided to demonstrate the data structure and allow for reproducibility checks.

Files

TB_Data_ML_Final for Zenodo.csv

Files (2.6 MB)

Name Size
md5:7ed0b0ff4a20c10f544a725e6cc95889
2.6 MB Preview Download
md5:d71aeba335dac604fc4ec6c5d6f5d404
14.5 kB Preview Download