Research on aging-friendly design risk assessment model based on non-parametric estimation.

Kernel density estimation was used to construct interaction behavior risk probability distributions for elderly users, capturing nonlinear fluctuations and long-tail extreme risks that mean-based methods fail to detect.

• Traditional mean-based indicators fail to capture elderly users' behavioral nonlinear fluctuations, long-tail extreme risks, and temporal mutations.
• Kernel density estimation (KDE) was applied to model the risk probability distributions of interaction behaviors.
• The non-parametric approach avoids assumptions about underlying data distributions, making it more suitable for high-uncertainty user groups.
• The model was empirically tested on 20 elderly participants using an induction cooker's intelligent touch interface.

Wavelet transform analysis was used to identify temporal risk evolution in elderly users' interaction behavior.

• Wavelet analysis was integrated into the model to detect temporal mutations in interaction behavior risk.
• This component specifically addresses the identification of risk changes over time that static or mean-based models cannot capture.
• The wavelet transform was combined with kernel density estimation and Value-at-Risk methods in the overall framework.

Value-at-Risk (VaR) was applied to establish prudent boundaries for key aging-friendly design parameters including size, spacing, font, and feedback rhythm.

• VaR was used to define prudent boundaries in the design parameter space.
• Key parameters assessed included button width, font size, and response delay.
• Monte Carlo simulation was used alongside VaR to minimize risk across the design parameter space.
• The model provides a closed-loop process described as 'behavior modeling → risk identification → parameter feedback'.

The model recommended specific design parameter thresholds that significantly reduce misoperation rates and task completion time for elderly users.

• Recommended parameters include button width ≥ 16 mm, font size ≥ 14pt, and response delay ≤ 500 ms.
• These parameters were identified through empirical testing with 20 elderly participants on an induction cooker's intelligent touch interface.
• Interaction data were collected across multiple parameter combinations.
• Applying the recommended parameters significantly reduced misoperation rates and task time while improving user satisfaction and safety.

The proposed model outperformed traditional mean-based assessment methods in capturing behavioral trends and controlling extreme-risk events for elderly users.

• Compared to traditional mean-based methods, the proposed model better captures behavioral trends and controls extremes.
• The model is described as suiting 'high-uncertainty user groups' such as the elderly.
• The model effectively identifies high-risk design zones that conventional approaches miss.
• The authors conclude the approach expands 'data-driven, prudent risk control in design science'.

The integrated risk assessment framework provides a closed-loop methodology applicable to multimodal interactions and intelligent elderly care beyond the tested induction cooker interface.

• The model establishes a 'behavior modeling → risk identification → parameter feedback' closed-loop for aging-friendly design.
• The authors note its value extends to multimodal interactions and intelligent elderly care applications.
• The study expands data-driven, prudent risk control methods within design science.
• The empirical validation was conducted specifically on an induction cooker's intelligent touch interface with 20 elderly participants.