Predictive Modeling of Menstrual Cycle Length: A Time Series Forecasting Approach

TL;DR

This paper explores machine learning and time series forecasting methods to accurately predict menstrual cycle length, aiding women's health management and planning.

Contribution

It introduces the application of various machine learning algorithms, including LSTM, for menstrual cycle prediction and demonstrates their effectiveness with synthetic data.

Findings

Accurate prediction of cycle onset and duration is feasible.

Time series models outperform traditional methods.

Synthetic data can effectively support model training.

Abstract

A proper forecast of the menstrual cycle is meaningful for women's health, as it allows individuals to take preventive actions to minimize cycle-associated discomforts. In addition, precise prediction can be useful for planning important events in a woman's life, such as family planning. In this work, we explored the use of machine learning techniques to predict regular and irregular menstrual cycles. We implemented some time series forecasting algorithm approaches, such as AutoRegressive Integrated Moving Average, Huber Regression, Lasso Regression, Orthogonal Matching Pursuit, and Long Short-Term Memory Network. Moreover, we generated synthetic data to achieve our purposes. The results showed that it is possible to accurately predict the onset and duration of menstrual cycles using machine learning techniques.