Lecture 5 Cheat Sheet – Non-Singleton & Type-2
Fuzzy Systems
Singleton Fuzzification
• Definition: Uses a crisp input directly (point value) for fuzzification.
• Analogy: Measuring room temperature as exactly 22°C and applying it directly to the fuzzy set.
Non-Singleton Fuzzification
• Definition: Uses uncertain (spread-out) input instead of a single crisp value, better handling
noisy/imprecise data.
• Analogy: A faulty thermometer says today’s temperature is somewhere between 21°C–23°C
instead of exactly 22°C.
Type-1 Fuzzy Set
• Definition: A fuzzy set where membership functions assign crisp degrees between 0 and 1.
• Analogy: A person’s height: 170 cm → 0.6 'tall,' 0.4 'medium'.
Type-2 Fuzzy Set
• Definition: Extends Type-1 by making the membership grade itself fuzzy (uncertain).
• Analogy: Two people disagree: one says 170 cm is 0.6 'tall,' another says 0.8. Type-2 stores
the whole range [0.6, 0.8].
Footprint of Uncertainty (FOU)
• Definition: The shaded region that captures all possible membership values in a Type-2 fuzzy
set.
• Analogy: Like weather forecasts: instead of one prediction (25°C), you keep a band (23–27°C)
to reflect uncertainty.
Interval Type-2 Fuzzy Sets
• Definition: A simplified form of Type-2 fuzzy sets where membership grades are intervals.
• Analogy: Saying 'The temperature feels warm between 40–60% confidence' instead of giving
an exact value.
Type Reduction
• Definition: The process of converting a Type-2 fuzzy set into a Type-1 fuzzy set before
defuzzification.
• Analogy: Like averaging multiple friends’ uncertain advice into one reasonable opinion before
making a decision.
Defuzzification
• Definition: Converting the reduced fuzzy set into a single crisp output.
• Analogy: After narrowing forecasts (23–27°C) into 25°C, you decide to dress for 25°C.
Applications
• Definition: Type-2 fuzzy systems are powerful in noisy, uncertain environments (speech
recognition, sensors, control systems).
• Analogy: Voice assistants (like Alexa/Siri) understanding speech despite background noise.
