Researchers at International Islamic University Islamabad develop a fuzzy rough aggregation approach combined with the MABAC multi-criteria decision method to evaluate and rank AI assistive technologies for disability support, handling uncertainty in performance criteria for more accurate tool selection.
Key points
Development of fuzzy rough Maclaurin symmetric mean (FRMSM) and its weighted dual variants for aggregation under uncertainty
Integration of FRMSM operators into the MABAC border approximation area method for multi-criteria decision-making
Application to classify and rank 10 AI assistive technologies, demonstrating improved selection accuracy for disability support
Why it matters:
This framework advances AI decision support by effectively handling uncertainty and interdependent criteria, improving assistive technology selection for disability care.
Q&A
What is a fuzzy rough set?
How does the MABAC method work?
What are Maclaurin symmetric mean aggregation operators?
How is this applied to AI assistive technology selection?
Read full article
Academy
Understanding Fuzzy Rough Sets and MABAC for AI Assistive Technology
Fuzzy Sets Overview: In traditional mathematics, an element either belongs or does not belong to a set. Fuzzy sets introduce a membership grade between 0 and 1, allowing partial belonging. For example, if a device’s accuracy is “good,” we might assign it a membership of 0.8 to the “high accuracy” set.
Rough Sets Overview: Rough set theory handles imprecise boundaries by defining lower and upper approximations. The lower approximation includes elements definitely in a set, while the upper approximation includes elements possibly in the set. Anything between these approximations is uncertain.
Fuzzy Rough Sets: By combining fuzzy membership with rough approximations, fuzzy rough sets capture both graded membership and uncertainty in classification. Each item is represented by a pair: its lower and upper fuzzy membership values. This hybrid model is ideal when expert opinions vary or criteria interact.
Maclaurin Symmetric Mean Aggregation
The Maclaurin Symmetric Mean (MSM) aggregates multiple inputs by averaging products of subsets of a given size and taking appropriate roots. This symmetric formulation makes the operator less sensitive to extremes. When extended to fuzzy rough numbers, the operator aggregates lower approximations separately from upper approximations, preserving uncertainty bounds.
Multi-criteria Decision Making (MCDM)
In many real-world decisions, multiple criteria must be considered simultaneously—for example, accuracy, ease of use, and adaptability of AI assistive devices. MCDM provides structured methods to compare alternatives across criteria, assign weights, and rank options.
The MABAC Method
- Normalization: Convert all criteria scores to a common scale (e.g., membership pairs for fuzzy rough sets).
- Weighting: Apply weights to each criterion to reflect its relative importance.
- Border Approximation Area (BAA): For each criterion, use an aggregation operator (such as fuzzy rough MSM) to compute the BAA, representing an average performance boundary.
- Distance Calculation: Measure how far each alternative’s weighted score is from the BAA. Positive distances indicate better-than-average performance.
- Ranking: Sum distances across criteria and rank alternatives by total score.
Why This Matters for AI Assistive Technology
AI-powered assistive devices—such as speech recognition systems, exoskeletons, or predictive analytics tools—offer personalized support for people with disabilities. However, selecting the optimal device requires balancing multiple, sometimes conflicting criteria under uncertainty. The fuzzy rough MABAC approach systematically handles data vagueness and interdependencies, guiding stakeholders to more reliable device choices and improving user outcomes.
