PREDICTION OF MENSTRUAL CYCLES WITH MACHINE LEARNING TECHNIQUES: AN APPLICATION WITH SYNTHETIC DATA

Abstract

Menstrual health is a major public health issue that affects millions of females in the world. Despite its importance, there is still a lack of accurate and reliable tools for predicting menstrual cycles. To address this gap, we have developed a machine learning model that utilises a large and diverse dataset of synthetic menstrual cycle data. In this paper, we have described the design and performance evaluation of the machine learning model in relation to average absolute deviation(MAE), average squared deviation (MSE), and accuracy. Our machine learning model is able to accurately predict menstrual cycles with an average absolute deviation(MAE) of less than 1 day, and a mean squared error (MSE) of less than 2 days. Our results demonstrate the potential of synthetic data generation techniques as a viable and cost-effective alternative to real-world data collection and storage and highlight the importance of careful design and validation of machine learning models for real-world applications.