Evaluating Interactive 2D Visualization as a Sample Selection Strategy for Biomedical Time-Series Data Annotation

TL;DR

This study compares sample selection methods for annotating biomedical time-series data, finding that interactive 2D visualizations improve annotation quality but carry higher risks and variability, especially with limited budgets.

Contribution

It introduces a graphical user interface-based 2D visualization method for sample selection and evaluates its effectiveness against traditional methods in real-world biomedical annotation tasks.

Findings

2D visualization outperforms other methods in aggregating labels across annotators.

2D visualization best captures rare classes in infant motility assessment.

Farthest-first traversal performs better with limited annotation budgets and individual annotators.

Abstract

Reliable machine-learning models in biomedical settings depend on accurate labels, yet annotating biomedical time-series data remains challenging. Algorithmic sample selection may support annotation, but evidence from studies involving real human annotators is scarce. Consequently, we compare three sample selection methods for annotation: random sampling (RND), farthest-first traversal (FAFT), and a graphical user interface-based method enabling exploration of complementary 2D visualizations (2DVs) of high-dimensional data. We evaluated the methods across four classification tasks in infant motility assessment (IMA) and speech emotion recognition (SER). Twelve annotators, categorized as experts or non-experts, performed data annotation under a limited annotation budget, and post-annotation experiments were conducted to evaluate the sampling methods. Across all classification tasks, 2DV performed best when aggregating labels across annotators. In IMA, 2DV most effectively captured rare classes, but also exhibited greater annotator-to-annotator label distribution variability resulting from the limited annotation budget, decreasing classification performance when models were trained on individual annotators' labels; in these cases, FAFT excelled. For SER, 2DV outperformed the other methods among expert annotators and matched their performance for non-experts in the individual-annotator setting. A failure risk analysis revealed that RND was the safest choice when annotator count or annotator expertise was uncertain, whereas 2DV had the highest risk due to its greater label distribution variability. Furthermore, post-experiment interviews indicated that 2DV made the annotation task more interesting and enjoyable. Overall, 2DV-based sampling appears promising for biomedical time-series data annotation, particularly when the annotation budget is not highly constrained.