Files
temperature-based-fertility…/thesis/sections/related_work.tex
T
2025-07-01 17:28:58 +02:00

80 lines
5.2 KiB
TeX
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
%! Author = alex
%! Date = 3/6/25
\section{Related Work}\label{sec:related_work}
There has been a variety of works in menstrual cycle analysis.
\subsection{Body Temperature}
\citeauthor{luo_detection_2020} used an in-ear wearable device that measured ear canal temperature every five minutes during sleep~\cite{luo_detection_2020}.
They trained a Hidden Markov Model (HMM) to classify each data point into either a high- or low-temperature state,
augmented with biorhythm information from the user.
After filtering, the final dataset consisted of 64 cycles, each with at least 40\% data availability and at least one self-reported ovulation day, as determined by a hormone test kit.
However, no information was provided regarding the distribution of cycle lengths or ovulation timing.
Ovulation detection was considered successful if the predicted day fell within ±3 days of the self-reported value.
The method achieved a sensitivity of 92.31\%, with 54.69\% of ovulation days detected exactly on the reported date.
The model, however, relies on strong assumptions of phase regularity and fixed transition durations—such as a standard
luteal phase length of 14 days—which do not reflect real-world variability.
Moreover, HMM predictions are conditioned on either a previous cycle or population-level averages, limiting performance in irregular or anovulatory cycles.
As a result, the approach performs well on regular, well-behaved data but lacks robustness in more diverse, real-world scenarios.
In~\citeyear{yu_tracking_2022}, \citeauthor{yu_tracking_2022} employed an in-ear thermometer along with a fitness tracker
for heart rate monitoring to predict the fertile window using machine learning~\cite{yu_tracking_2022}.
Their study population consisted of 153 women, divided into a regular cycle group ($n = 103$) and an irregular group ($n = 50$).
After filtering, 89 and 25 participants remained in the regular and irregular groups, respectively.
The prediction task was to determine whether a given day falls within the fertile window, based on data from the preceding days.
A second model was trained to predict whether menstruation occurs on a given day, again using preceding data as input.
For the fertile window prediction, the model achieved a sensitivity of 69.30\% in the regular group and 21.00\% in the irregular group.
For menstruation prediction, the model detected 70.70\% of menstruation days in the regular group and 36.30\% in the irregular group.
These results indicate that the model performs reasonably well for individuals with regular cycles,
but struggles significantly in the presence of menstrual irregularity—particularly in detecting the fertile window.
Several studies have questioned the utility of BBT (Basal Body Temperature) for reliable ovulation prediction.
For example \citeauthor{bauman_basal_1981} concluded, that BBT is not a robust standalone marker due to its
retrospective nature and sensitivity to external factors and thus must be used with extreme caution clinical or research evaluations~\cite{bauman_basal_1981}
\subsection{Alternative Pyhysiological Signals}
In addition to temperature, other physiological signals have been explored for ovulation and cycle phase prediction.
For example,\citeauthor{masuda_machine_2025} developed a machine learning algorithm to classify phases of the menstrual cycle
(follicular vs. luteal) based on sleeping heart rate, as recorded by a fitness tracker~\cite{masuda_machine_2025}.
They used an XGBoost classifier for this binary task and additionally performed ovulation day prediction,
although the details of this task were not fully specified.
Ground truth labels were derived from self-reported data and daily ovulation predictor kits.
The dataset used in their study was restricted to healthy, non-pregnant women aged 1834 with natural menstrual cycles.
Anovulatory cycles were excluded, along with users meeting the following criteria:
\begin{itemize}
\item Use of medications, including hormonal contraceptives
\item Travel across time zones
\item Night-shift work
\item Pregnancy within the past year
\item Sleep disorders
\item Tobacco use
\end{itemize}
They further subdivided participants into groups with low and high sleep variability, called HVST and LVST respectively.
As a result, the study population and cycle types were highly regular and homogeneous,
with 30 cycles (18 women) in the HVST and 26 cycles (16 women) in the LVST category.
No information about the distribution of both lengths or ovulation dates was given.
Based on selected features—such as minimum sleeping heart rate and single-point basal body temperature (BBT) after
waking—they report classification accuracies between 0.843 and 0.864, depending on the feature subset,
with very similar numbers for precision, recall, specificity and F1 score.
Ovulation day prediction yielded an average absolute error between 3.6 and 4.1 days.
When models are trained on highly constrained datasets with predictable patterns and clear ovulatory signals,
complex methods often show limited gains over naive or rule-based approaches—as will be demonstrated in Section~\ref{sec:methodology}.
%parts: traditional methods, ml approaches, tft in medical time series or other time series