Interval Type-2 Fuzzy Sets: Behaviour and Performance
Comparative study of type-1, general type-2, and interval type-2 fuzzy sets, introducing the Interval Creation Approach.
University of Dhaka · Supervisor: Dr. Shaily Kabir (LUCID Lab (Lab for Uncertainty in Data and Decision Making), University of Nottingham) · March 2023 – January 2024
Overview
Investigated the comparative efficiency of type-1, general type-2, and interval type-2 fuzzy sets, and designed the Interval Creation Approach (ICA) — a methodology for constructing interval type-2 fuzzy sets from a range of type-1 fuzzy set shapes: triangular, trapezoidal, Gaussian, generalized bell-shaped, and both symmetric and asymmetric curves.
Approach
1
Methodology design
Designed the complete ICA methodology to construct interval type-2 fuzzy sets from the full range of type-1 fuzzy set shapes.
2
Evaluation planning
Built a work-plan to demonstrate ICA’s effectiveness using both synthetic and real-world datasets.
3
Benchmarking
Performed a detailed comparison of ICA against the Interval Approach (IA), Enhanced Interval Approach (EIA), and M2 methods — the previously established approaches for this conversion.
4
Efficiency gains
Reduced extra pre-processing time compared to prior methods, improving overall computational cost.
Outcomes
- First-author paper: “An Interval Creation Approach to Construct Interval Type-2 Fuzzy Sets”, published at the 2024 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), part of the 2024 IEEE World Congress on Computational Intelligence (WCCI)
- Presented in the session “Virtual: Foundations of Fuzzy Systems” at FUZZ-IEEE 2024, on June 30, 2024 — presentation slides
See publications for full citation details.
Skills
Python Java C++ MATLAB PyIT2FLS Juzzy Collaborative & content-based filtering Neural networks Evolutionary algorithms Presentation skills Paper & report writing