![[2602.22400] Predicting Multi-Drug Resistance in Bacterial Isolates Through Performance Comparison and LIME-based Interpretation of Classification Models](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22400] Predicting Multi-Drug Resistance in Bacterial Isolates Through Performance Comparison and LIME-based Interpretation of Classification Models
Summary
This study presents a machine learning framework to predict multi-drug resistance (MDR) in bacterial isolates, utilizing five classification models and LIME for interpretability.
Why It Matters
The rise of antimicrobial resistance poses significant challenges in clinical settings. This research enhances predictive capabilities for MDR, aiding in timely clinical decision-making and improving patient outcomes through better antimicrobial stewardship.
Key Takeaways
- Five classification models were evaluated for predicting MDR, with XGBoost and LightGBM showing the best performance.
- The study utilized a dataset of 9,714 bacterial isolates to train the models, focusing on antibiotic resistance patterns.
- LIME was applied to enhance model interpretability, identifying key resistance contributors.
- The framework supports earlier identification of MDR, crucial for effective clinical interventions.
- Combining high accuracy with interpretability fosters trust in machine learning applications in healthcare.
Computer Science > Machine Learning arXiv:2602.22400 (cs) [Submitted on 25 Feb 2026] Title:Predicting Multi-Drug Resistance in Bacterial Isolates Through Performance Comparison and LIME-based Interpretation of Classification Models Authors:Santanam Wishal, Riad Sahara View a PDF of the paper titled Predicting Multi-Drug Resistance in Bacterial Isolates Through Performance Comparison and LIME-based Interpretation of Classification Models, by Santanam Wishal and 1 other authors View PDF Abstract:The rise of Antimicrobial Resistance, particularly Multi-Drug Resistance (MDR), presents a critical challenge for clinical decision-making due to limited treatment options and delays in conventional susceptibility testing. This study proposes an interpretable machine learning framework to predict MDR in bacterial isolates using clinical features and antibiotic susceptibility patterns. Five classification models were evaluated, including Logistic Regression, Random Forest, AdaBoost, XGBoost, and LightGBM. The models were trained on a curated dataset of 9,714 isolates, with resistance encoded at the antibiotic family level to capture cross-class resistance patterns consistent with MDR definitions. Performance assessment included accuracy, F1-score, AUC-ROC, and Matthews Correlation Coefficient. Ensemble models, particularly XGBoost and LightGBM, demonstrated superior predictive capability across all metrics. To address the clinical transparency gap, Local Interpretable Model-agnostic E...
