Unboxing Industry-Standard AI Models for Male Fertility Prediction with SHAP.

Area of Science:

• Reproductive Medicine
• Artificial Intelligence in Healthcare
• Bioinformatics

Background:

• Male infertility affects approximately 30% of couples and is often underrecognized.
• Lifestyle and environmental factors contribute to declining male fertility rates globally.
• Early detection of male infertility is crucial for effective management and treatment.

Purpose of the Study:

• To compare the performance of seven machine learning (ML) models for detecting male infertility.
• To identify the most effective ML model for male fertility classification.
• To utilize Shapley Additive Explanations (SHAP) for model interpretability.

Main Methods:

• Implementation of seven standard ML models: support vector machine, random forest, decision tree, logistic regression, naïve bayes, adaboost, and multi-layer perception.
• Utilizing five-fold cross-validation on a balanced dataset.
• Applying SHAP for feature importance analysis and model explanation.

Main Results:

• The Random Forest (RF) model achieved the highest accuracy (90.47%) and Area Under the Curve (AUC) (99.98%).
• SHAP analysis provided insights into the decision-making processes of both high-performing and low-performing models.
• Comparative analysis identified optimal ML models for male fertility detection.

Conclusions:

• ML models, particularly Random Forest, offer a robust approach for early male infertility detection.
• SHAP explanations enhance the interpretability of ML models, aiding clinical decision-making.
• This study provides a reference for clinicians in planning male infertility treatments using AI.