Artificial Intelligence & Machine Learning
Course Code : 4803 & 4829
Group-3
Submitted to:
Prof. Engr. Md. Razu Ahmed
Professor & Chairman
Dept. of CCE, IIUC
Submitted by:
Syed Md Azizul Alam (E211016)
Mehraj Ibne Halim (E211018)
Al Mahmud Arafat (E211019R)
Fuzzy Logic
Traditional Logic (Classical/Binary Logic):
• Based on two-valued logic: Either 0 or 1
• Used in most digital circuits, computers, and decision-making
• Works well when the data is clear-cut (ex, ON/OFF, TRUE/FALSE)
Example:
• Temperature > 30°C Hot (1)
→
• Temperature 30°C Not Hot (0)
≤ →
• it's either hot or not.
Fuzzy Logic:
• Allows intermediate values between 0 and 1 (like 0.2, 0.5, 0.9 etc.)
• Captures uncertainty, like real-life situations
• Instead of absolute truth, it assigns degrees of truth
Example:
• Temperature = 32°C
• 0.7 Hot, 0.3 Warm
• Not fully hot, not fully warm — it’s partially both
Fuzzy Logic
Fuzzy Sets: A set where elements have degrees of membership.
Membership Function (MF): Maps input to a value between 0 and 1.
Variables like “Temperature” with values like Cold, Warm, Hot.
Fuzzification: Converts crisp input to fuzzy values using MF.
Defuzzification: Converts fuzzy output back to crisp values.
50 70 90 110
30
10
Temp. (F°)
Freezing Cool Warm Hot
0
1
Fuzzy Logic
Fuzzification Process:
Input: 36°F
Output: 70% Freezing, 30% Cold
50 70 90 110
30
10
Temp. (F°)
Freezing Cool Warm Hot
0
1
Fuzzy Inference System (FIS)
A Fuzzy Inference System (FIS) uses human-like reasoning to make smart decisions. It
works in five steps, as shown in the diagram.
1. Input (Crisp): The system receives a clear input value (example: 35°C temperature).
2. Fuzzification Interface: This part converts crisp input into fuzzy terms, like “hot” or
“warm”. Now the system can think like humans.
Fuzzy Inference System (FIS)
3. Knowledge Base:
• Database: Defines fuzzy sets and terms.
• Rule Base: Contains IF–THEN rules like: IF temperature is hot AND fan is slow THEN increase
speed.
4. Decision-Making Unit (Inference Engine): This is the heart of FIS. It checks the rules and fuzzy
inputs, and decides the fuzzy output.
5. Defuzzification Interface: This part converts fuzzy output back to a crisp value, like “increase fan
to 75% speed”.
6. Final Output (Crisp): A clear action is given — easy for machines to understand and apply.
Method of FIS
Mamdani Fuzzy Inference System
How It Works:
1. Fuzzification: Converts inputs into fuzzy
values (ex, Low, Medium, High).
2. Rule Application: Uses rules like “IF
temperature IS High, THEN fan speed IS
High”.
3. Aggregation: Combines all the rules' results.
4. Defuzzification: Converts fuzzy result back
into a crisp value.
Example: Used in control systems like air
conditioning.
Method of FIS
Takagi-Sugeno Fuzzy Inference System
How It Works:
1. Fuzzification: Converts inputs into fuzzy values.
2. Rule Application: Uses rules like “IF temperature
IS High, THEN fan speed = a * temperature + b”.
3. Aggregation: Combines the results with
weighted averages.
4. Defuzzification: Converts the final fuzzy result to
a crisp value.
Example: Used in more precise calculations or
optimizations.

Artificial intelligence and machine learning.pptx

  • 1.
    Artificial Intelligence &Machine Learning Course Code : 4803 & 4829 Group-3 Submitted to: Prof. Engr. Md. Razu Ahmed Professor & Chairman Dept. of CCE, IIUC Submitted by: Syed Md Azizul Alam (E211016) Mehraj Ibne Halim (E211018) Al Mahmud Arafat (E211019R)
  • 2.
    Fuzzy Logic Traditional Logic(Classical/Binary Logic): • Based on two-valued logic: Either 0 or 1 • Used in most digital circuits, computers, and decision-making • Works well when the data is clear-cut (ex, ON/OFF, TRUE/FALSE) Example: • Temperature > 30°C Hot (1) → • Temperature 30°C Not Hot (0) ≤ → • it's either hot or not. Fuzzy Logic: • Allows intermediate values between 0 and 1 (like 0.2, 0.5, 0.9 etc.) • Captures uncertainty, like real-life situations • Instead of absolute truth, it assigns degrees of truth Example: • Temperature = 32°C • 0.7 Hot, 0.3 Warm • Not fully hot, not fully warm — it’s partially both
  • 3.
    Fuzzy Logic Fuzzy Sets:A set where elements have degrees of membership. Membership Function (MF): Maps input to a value between 0 and 1. Variables like “Temperature” with values like Cold, Warm, Hot. Fuzzification: Converts crisp input to fuzzy values using MF. Defuzzification: Converts fuzzy output back to crisp values. 50 70 90 110 30 10 Temp. (F°) Freezing Cool Warm Hot 0 1
  • 4.
    Fuzzy Logic Fuzzification Process: Input:36°F Output: 70% Freezing, 30% Cold 50 70 90 110 30 10 Temp. (F°) Freezing Cool Warm Hot 0 1
  • 5.
    Fuzzy Inference System(FIS) A Fuzzy Inference System (FIS) uses human-like reasoning to make smart decisions. It works in five steps, as shown in the diagram. 1. Input (Crisp): The system receives a clear input value (example: 35°C temperature). 2. Fuzzification Interface: This part converts crisp input into fuzzy terms, like “hot” or “warm”. Now the system can think like humans.
  • 6.
    Fuzzy Inference System(FIS) 3. Knowledge Base: • Database: Defines fuzzy sets and terms. • Rule Base: Contains IF–THEN rules like: IF temperature is hot AND fan is slow THEN increase speed. 4. Decision-Making Unit (Inference Engine): This is the heart of FIS. It checks the rules and fuzzy inputs, and decides the fuzzy output. 5. Defuzzification Interface: This part converts fuzzy output back to a crisp value, like “increase fan to 75% speed”. 6. Final Output (Crisp): A clear action is given — easy for machines to understand and apply.
  • 7.
    Method of FIS MamdaniFuzzy Inference System How It Works: 1. Fuzzification: Converts inputs into fuzzy values (ex, Low, Medium, High). 2. Rule Application: Uses rules like “IF temperature IS High, THEN fan speed IS High”. 3. Aggregation: Combines all the rules' results. 4. Defuzzification: Converts fuzzy result back into a crisp value. Example: Used in control systems like air conditioning.
  • 8.
    Method of FIS Takagi-SugenoFuzzy Inference System How It Works: 1. Fuzzification: Converts inputs into fuzzy values. 2. Rule Application: Uses rules like “IF temperature IS High, THEN fan speed = a * temperature + b”. 3. Aggregation: Combines the results with weighted averages. 4. Defuzzification: Converts the final fuzzy result to a crisp value. Example: Used in more precise calculations or optimizations.