Slovenijo v informacijsko drubo

1
(No Transcript)
2
Subfields

• Bayes rules
• Expert systems
• Hybrid systems
• Rules
• Agents
• Learning

3
Rules

• If condition1 and condition2 then Consequence
• Inference engine, lists of rules (knowledge
  base), data (database), knowledge encapsulation,
  explanation, user
• Rule chaning foreward, backward, combinations
• Deduction, abduction, induction
• Semantic networks, objects, frames, trees,
  ontologies

4
some concepts

• Declarative-procedural programming
• Search depth-first, breadth-first, beam-search,
  alfa-beta, A
• Meta-
• Progaming languages Prolog, Logo, Lisp ...
• Logic closed world

5
Probabilistic-Bayesian

• Optimal in theory, in future not
• P(HE) P(H)xP(EH) / P(E)
• P(H)xP(EH)
• P(HE) --------------------------------------
• P(H)xP(EH) P(H)xP(EH)
• Likehood ratio, odds
• O(H) P(H)/ P(H)

6
Naïve Baesian Classifier

• P(A1v1A2v2Cc)
  
• P(CA) P(Cc) x ---------------------------------
  
• P(A1v1A2v2)
• P(A1v1Cc)xP(A2v2Cc)

7
Baesian Network
P(F) P(FL,A)P(L?A) P(FL,A) P(L?A)
P(FL,A) P(L?A) P(FL,A) P(L?A)