Researchers present a systematic empirical evaluation of learning curve convergence rates to address the lack of data-driven guidelines for determining minimum sample sizes in inertial sensor-based classification tasks.
- The study introduces a unified framework analyzing classification performance under binary and multi-class scenarios, deriving an empirical formula to estimate performance relative to dataset size.
- Testing across six diverse real-world datasets totaling 102.7 hours of inertial measurements demonstrates that accuracy follows a consistent logarithmic growth pattern regardless of task complexity.
- The authors propose a quantitative stability point metric defined as the sample size required for the learning curve to stabilize within a predefined mean absolute percentage deviation of its asymptotic maximum.
- Analysis reveals that models often reach practical stability with substantially fewer samples than traditional heuristics suggest, allowing extrapolation of total data requirements from small-scale pilot studies.
These findings offer concrete, data-backed guidelines for planning recording campaigns in inertial sensing applications, shifting the paradigm from maximizing data volume toward optimizing data efficiency.