Hacker, speaker, entrepreneur
What... is it?
How... does it work?
Why... would I use it?
- The Logic ... (explicit rules defined by humans)
- ... of Fuzzy sets ... (group membership is not clear-cut)
- ... that can be optimized (improved with usage)
- Have degrees of membership (0.0 -> 1.0)
- Represent vague terms from the real world
- Can overlap
Age (in years)
Another fuzzy set
Savings rate (percentage of income)
- Rules statements
- About fuzzy sets
- Implications between them
IF Age IS Young THEN Savings Rate IS High.
IF Age IS Old THEN Savings Rate IS Low.
Notes about rules
- "Then" part fires proportionate to the "If" part
- Multiple rules can fire for the same input
- Many inputs can be used (e.g., income, # of children)
- Rules can be weighted (some rules are more important)
IF Age is Young AND Number of Children is Low Then Savings Rate is High.
IF Age is Old OR Income is Low Then Savings Rate is Low.
IF NOT Number of Children is Low Then Savings Rate is Low.
- Improve the parameters to give better results
- Direct results ("That's not the right savings rate!")
- Indirect results ("I had enough savings to buy a boat!")
Why use it?
- When explicit rules matter, but you want optimization, too
- When there is a lot of organizational knowledge
- When you don't have a lot of training data for a pure learning system
- When terms in the problem domain are vague
Understanding Fuzzy Logic
By Evan Prodromou