Clinical Pregnancy Prediction Models for Intrauterine Insemination

957 intrauterine insemination cycles were used to develop and evaluate three machine learning models for predicting clinical pregnancy. Yingwei Fu authored this research, which compares logistic regression, random forest, and multilayer perceptron algorithms. The dataset was last updated on April 24, 2026.

Use Cases

• Compare predictive performance of logistic regression, random forest, and multilayer perceptron models based on the described methodology.
• Explore clinical interpretability of fertility prediction models using SHAP analysis based on identified important predictors.
• Address class imbalance in medical datasets using Synthetic Minority Oversampling Technique (SMOTE) based on the described preprocessing.
• Perform risk stratification for assisted reproduction procedures based on the model's predictive features.
• Evaluate model stability across folds for clinical applications based on the reported cross-validation results.

Strengths

• The dataset contains 957 IUI cycles, providing a specific sample size for model development.
• Model performance was evaluated using 10-fold cross-validation, a robust methodological approach.
• SHAP analysis identified six important clinical predictors, offering interpretability insights.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The predictive models demonstrated modest discriminative ability, with AUC values around 0.55.

Provenance

Source

figshare

Collection Method

Data likely collected from clinical IUI cycles for research purposes.

Freshness

Last updated 2026-04-24 04:23:28; freshness should be verified.