Auto machine learning to predict pregnancy after fresh embryo transfer following in vitro fertilization

Introduction: The use of human reproduction techniques (ART) to obtain pregnancy are increasing. However pregnancy rates after ART remain as low as around 30%. The use of machine Learning (ML) is increasing in medicine and prediction models are helpful to preview the outcome of in vitro fertilization(IVF) cycles.
Methods: Data from IVF cycles with fresh embryo transfer between January 2018 and December 2021 were collected. The Auto Machine Learning (Auto ML) PyCaret was used to construct the model and predict the clinical pregnancy rate.
Results: Among 14 ML algorithms, Ridge Classification (RC) has the best accuracy(57,69%). Transfer in day 5 was the most important feature related to the outcome.
Conclusion: Despite the low accuracy as a result of a small sample, familiarization with ML models, as well as awareness of the importance of data collection should be part of daily activities of physicians and healthcare professionals in the field of ART.