Abstract:INTRODUCTION: Assisted reproductive technologies (ART) are recent improvements in infertility treatment. However, there is no significant increase in pregnancy rates with the aid of ART. Costly and complex process of ART's makes them as challenging issues. Computational prediction models could predict treatment outcome, before the start of an ART cycle.AIM: This review provides an overview on machine learning-based prediction models in ART.METHODS: This article was executed based on a literature review through scientific databases search such as PubMed, Scopus, Web of Science and Google Scholar.RESULTS: We identified 20 papers reporting on machine learning-based prediction models in IVF or ICSI settings. All of the models were validated only by internal validation. Therefore, external validation of the models and the impact analysis of them were the missing parts of the all studies.CONCLUSION: Machine learning-based prediction models provide a clinical decision support tool for both clinicians and patients and lead to improvement in ART success rates.

A Review of Machine Learning Approaches in Assisted Reproductive Technologies