Automated Analysis

This demo uses Support Vector Machines (SVMs), a robust machine learning method, to analyse nocturnal pulse oximetry data in children. Computed statistical, signal processing, and oximetric features are used in the prediction process.

ML Sleep Staging of oximetry data is performed using an SVM Classifier (SVC). Thirty-second segments (epochs) of the recording are classified as Sleep or Wake, and these segments are filtered before analysis of oximetry data for sleep-disordered breathing. Sleep staging enables the computation of sleep statistics traditionally obtained from overnight polysomnography, such as sleep duration, latency, efficiency, and WASO. While this sleep staging step can be disabled, it is recommended for analysing oximetry data collected in uncontrolled environments (e.g., at home).

Analysis of recording to uses either a SVM Regressor (SVR) to produce a point estimate of the apnoea-hypopnea index (AHI) with accompanying uncertainty bounds, or an SVC to predict whether the apnoea-hypopnea index (AHI) is ≥5.

Peer-reviewed article containing out-of-sample performance data coming soon.

Online Retraining

This demo also offers a streamlined pipeline retraining and validating models using a bank of pre-computed features and datasets. Newly trained models are immediately available for use after saving, enabling rapid prototyping and rollout of updated models. New data can be added to existing models and the demo allows for warm-starting existing models with these new data to substantially reducing training time.

A custom solver using Sequential Minimal Optimisation (Platt, 1998) and a set of published heuristics (Fan et al., 2005) is used. Performance of trained models using identical data and hyper-parameters are performance-equivalent to models produced by other commonly used tools.