SlideShare una empresa de Scribd logo
1 de 23
MUSA AL HAWAMDAH / 128129001011
          15-10-2012
Concept Learning as Search:
 Concept learning can be viewed as the task of
 searching through a large space of hypothesis
 implicitly defined by the hypothesis representation.

 The goal of the concept learning search is to find the
 hypothesis that best fits the training examples.
General-to-Specific LearningMost
General-to-Specific Learning:
Example   Sky     AirTemp    Humidity    Wind       Water      Forecast   EnjoySport

   1      Sunny    Warm       Normal     Strong     Warm        Same         Yes

   2      Sunny    Warm        High      Strong     Warm        Same         Yes

   3      Rainy     Cold       High      Strong     Warm       Change        No

   4      Sunny    Warm        High      Strong         Cool   Change        Yes




                            h1=(Sunny,?,?,Strong,?,?)
                              h2=(Sunny,?,?,?,?,?)

  *h2 is more general than h1.
  *h2 imposes fewer constraints on the instance than h1.
FIND-S: Finding a Maximally
Specific Hypothesis:
1-Initialize h to the most specific hypotesis in h
2-for each positive training instance x
             - For each attribute constraint aj in h
                   if the constraint aj is satisfied by x
                   then do nothing
                    else replace aj in h by the next more
  general constraint that is satisfied by x

3- output hypothesis h.
Step 1: FIND-S:
Example      Sky      AirTemp    Humidity     Wind     Water   Forecast   EnjoySport

   1        Sunny      Warm       Normal      Strong   Warm     Same         Yes

   2        Sunny      Warm        High       Strong   Warm     Same         Yes

   3        Rainy       Cold       High       Strong   Warm    Change        No

   4        Sunny      Warm        High       Strong   Cool    Change        Yes




 Initialize h to the most specific hypotesis in h


 h0 = <Ø, Ø, Ø, Ø, Ø, Ø>
Step 2: FIND-S :
 Version Space:
    The set of all valid hypotheses provided by an algorithm is
    called version space (VS)with respect to the hypothesis
    space Hand the given example set D.



 Candidate-Elimination Algorithm:
* The Candidate-Eliminationalgorithm finds all
 describable hypotheses that are consistent with the
 observed training examples.
* Hypothesis is derived from examples regardless of
 whether x is positive or negative example
LIST-THEN-ELIMINATE Algorithm
to Obtain Version Space:
 In principle, the LIST-THEN-ELIMINATE algorithm
 can be applied whenever the hypothesis space H is
 finite.

 It is guaranteed to output all hypotheses consistent
 with the training data.

 Unfortunately, it requires exhaustively enumerating all
 hypotheses in H-an unrealistic requirement for all but
 the most trivial hypothesis spaces.
Candidate-Elimination Algorithm:
•The CANDIDATE-ELIMINATION algorithm works
   on the same principleas the above LIST-THEN-
   ELIMINATE algorithm.
•It employs a much more compact representation of
   the version space.
 •In this the version spaceis represented by its most
   general and least general members (Specific).
 •These members form general and specific
   boundary sets that delimit the version space
   within the partially ordered hypothesis space.
Example :
Example   Sky     AirTemp   Humidity   Wind     Water   Forecast   EnjoySport

   1      Sunny    Warm      Normal    Strong   Warm     Same         Yes

   2      Sunny    Warm       High     Strong   Warm     Same         Yes

   3      Rainy    Cold       High     Strong   Warm    Change        No

   4      Sunny    Warm       High     Strong   Cool    Change        Yes
What will Happen if the Training
Contains errors ?
Example   Sky     AirTemp   Humidity   Wind     Water   Forecast   EnjoySport

   1      Sunny    Warm      Normal    Strong   Warm     Same         Yes

   2      Sunny    Warm       High     Strong   Warm     Same         No

   3      Rainy    Cold       High     Strong   Warm    Change        No

   4      Sunny    Warm       High     Strong   Cool    Change        Yes
Concept learning
Concept learning
Concept learning

Más contenido relacionado

La actualidad más candente

Inductive bias
Inductive biasInductive bias
Inductive biasswapnac12
 
Back propagation
Back propagationBack propagation
Back propagationNagarajan
 
Multilayer & Back propagation algorithm
Multilayer & Back propagation algorithmMultilayer & Back propagation algorithm
Multilayer & Back propagation algorithmswapnac12
 
Logistic regression in Machine Learning
Logistic regression in Machine LearningLogistic regression in Machine Learning
Logistic regression in Machine LearningKuppusamy P
 
Issues in knowledge representation
Issues in knowledge representationIssues in knowledge representation
Issues in knowledge representationSravanthi Emani
 
Machine learning Lecture 2
Machine learning Lecture 2Machine learning Lecture 2
Machine learning Lecture 2Srinivasan R
 
I. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHMI. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHMvikas dhakane
 
ELEMENTS OF TRANSPORT PROTOCOL
ELEMENTS OF TRANSPORT PROTOCOLELEMENTS OF TRANSPORT PROTOCOL
ELEMENTS OF TRANSPORT PROTOCOLShashank Rustagi
 
Inductive analytical approaches to learning
Inductive analytical approaches to learningInductive analytical approaches to learning
Inductive analytical approaches to learningswapnac12
 
Machine Learning-Linear regression
Machine Learning-Linear regressionMachine Learning-Linear regression
Machine Learning-Linear regressionkishanthkumaar
 
3.2 partitioning methods
3.2 partitioning methods3.2 partitioning methods
3.2 partitioning methodsKrish_ver2
 
Control Strategies in AI
Control Strategies in AI Control Strategies in AI
Control Strategies in AI Bharat Bhushan
 
Transfer Learning: An overview
Transfer Learning: An overviewTransfer Learning: An overview
Transfer Learning: An overviewjins0618
 
Types of Machine Learning
Types of Machine LearningTypes of Machine Learning
Types of Machine LearningSamra Shahzadi
 
Machine Learning with Decision trees
Machine Learning with Decision treesMachine Learning with Decision trees
Machine Learning with Decision treesKnoldus Inc.
 
Performance analysis(Time & Space Complexity)
Performance analysis(Time & Space Complexity)Performance analysis(Time & Space Complexity)
Performance analysis(Time & Space Complexity)swapnac12
 
Peephole optimization techniques in compiler design
Peephole optimization techniques in compiler designPeephole optimization techniques in compiler design
Peephole optimization techniques in compiler designAnul Chaudhary
 
2.4 rule based classification
2.4 rule based classification2.4 rule based classification
2.4 rule based classificationKrish_ver2
 

La actualidad más candente (20)

Inductive bias
Inductive biasInductive bias
Inductive bias
 
Back propagation
Back propagationBack propagation
Back propagation
 
Multilayer & Back propagation algorithm
Multilayer & Back propagation algorithmMultilayer & Back propagation algorithm
Multilayer & Back propagation algorithm
 
Logistic regression in Machine Learning
Logistic regression in Machine LearningLogistic regression in Machine Learning
Logistic regression in Machine Learning
 
Issues in knowledge representation
Issues in knowledge representationIssues in knowledge representation
Issues in knowledge representation
 
Machine learning Lecture 2
Machine learning Lecture 2Machine learning Lecture 2
Machine learning Lecture 2
 
I. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHMI. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHM
 
ELEMENTS OF TRANSPORT PROTOCOL
ELEMENTS OF TRANSPORT PROTOCOLELEMENTS OF TRANSPORT PROTOCOL
ELEMENTS OF TRANSPORT PROTOCOL
 
Inductive analytical approaches to learning
Inductive analytical approaches to learningInductive analytical approaches to learning
Inductive analytical approaches to learning
 
Machine Learning-Linear regression
Machine Learning-Linear regressionMachine Learning-Linear regression
Machine Learning-Linear regression
 
3.2 partitioning methods
3.2 partitioning methods3.2 partitioning methods
3.2 partitioning methods
 
Control Strategies in AI
Control Strategies in AI Control Strategies in AI
Control Strategies in AI
 
Max net
Max netMax net
Max net
 
Transfer Learning: An overview
Transfer Learning: An overviewTransfer Learning: An overview
Transfer Learning: An overview
 
Clustering
ClusteringClustering
Clustering
 
Types of Machine Learning
Types of Machine LearningTypes of Machine Learning
Types of Machine Learning
 
Machine Learning with Decision trees
Machine Learning with Decision treesMachine Learning with Decision trees
Machine Learning with Decision trees
 
Performance analysis(Time & Space Complexity)
Performance analysis(Time & Space Complexity)Performance analysis(Time & Space Complexity)
Performance analysis(Time & Space Complexity)
 
Peephole optimization techniques in compiler design
Peephole optimization techniques in compiler designPeephole optimization techniques in compiler design
Peephole optimization techniques in compiler design
 
2.4 rule based classification
2.4 rule based classification2.4 rule based classification
2.4 rule based classification
 

Destacado

Laws of learning
Laws of learningLaws of learning
Laws of learningHans Mallen
 
Knowledge Construction Mindtools
Knowledge Construction MindtoolsKnowledge Construction Mindtools
Knowledge Construction Mindtoolstnwest
 
Constructivism : Knowledge Construction/Concept Learning
Constructivism : Knowledge Construction/Concept LearningConstructivism : Knowledge Construction/Concept Learning
Constructivism : Knowledge Construction/Concept LearningAl Francis Galabo
 
Constructivism: Knowledge Construction / Concept Learning
Constructivism: Knowledge Construction / Concept LearningConstructivism: Knowledge Construction / Concept Learning
Constructivism: Knowledge Construction / Concept LearningChoc Nat
 
Principles of learning (6 laws of learning)
Principles of learning (6 laws of learning)Principles of learning (6 laws of learning)
Principles of learning (6 laws of learning)Jen Chua
 
Psychology: Learning
Psychology: LearningPsychology: Learning
Psychology: Learninghm alumia
 
Concept learning presentation: overview
Concept learning presentation: overviewConcept learning presentation: overview
Concept learning presentation: overviewkitkeller
 
Principles of constructivism
Principles of constructivism Principles of constructivism
Principles of constructivism Elleh Rachelle
 
Constructivism ppt
Constructivism pptConstructivism ppt
Constructivism pptamysuesue
 
[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning
[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning
[Facilitating Learning] Constructivism: Knowledge Construction/Concept LearningJhoanna Lyn
 
The PRINCIPLES of LEARNING (Principles of Teaching 1)
The PRINCIPLES of LEARNING (Principles of Teaching 1)The PRINCIPLES of LEARNING (Principles of Teaching 1)
The PRINCIPLES of LEARNING (Principles of Teaching 1)Taguig City University
 
Laws of learning
Laws of learningLaws of learning
Laws of learningchezka8
 

Destacado (16)

Laws of learning
Laws of learningLaws of learning
Laws of learning
 
Knowledge Construction Mindtools
Knowledge Construction MindtoolsKnowledge Construction Mindtools
Knowledge Construction Mindtools
 
Week 4 Concepts of Learning
Week 4 Concepts of LearningWeek 4 Concepts of Learning
Week 4 Concepts of Learning
 
Constructivism : Knowledge Construction/Concept Learning
Constructivism : Knowledge Construction/Concept LearningConstructivism : Knowledge Construction/Concept Learning
Constructivism : Knowledge Construction/Concept Learning
 
Constructivism: Knowledge Construction / Concept Learning
Constructivism: Knowledge Construction / Concept LearningConstructivism: Knowledge Construction / Concept Learning
Constructivism: Knowledge Construction / Concept Learning
 
Principles of learning
Principles of learningPrinciples of learning
Principles of learning
 
Principles of learning (6 laws of learning)
Principles of learning (6 laws of learning)Principles of learning (6 laws of learning)
Principles of learning (6 laws of learning)
 
Psychology: Learning
Psychology: LearningPsychology: Learning
Psychology: Learning
 
Concept learning presentation: overview
Concept learning presentation: overviewConcept learning presentation: overview
Concept learning presentation: overview
 
Principles of constructivism
Principles of constructivism Principles of constructivism
Principles of constructivism
 
Constructivism ppt
Constructivism pptConstructivism ppt
Constructivism ppt
 
[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning
[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning
[Facilitating Learning] Constructivism: Knowledge Construction/Concept Learning
 
7 laws of learning
7 laws of learning7 laws of learning
7 laws of learning
 
The PRINCIPLES of LEARNING (Principles of Teaching 1)
The PRINCIPLES of LEARNING (Principles of Teaching 1)The PRINCIPLES of LEARNING (Principles of Teaching 1)
The PRINCIPLES of LEARNING (Principles of Teaching 1)
 
Laws of learning
Laws of learningLaws of learning
Laws of learning
 
Learning ppt
Learning ppt Learning ppt
Learning ppt
 

Concept learning

  • 1. MUSA AL HAWAMDAH / 128129001011 15-10-2012
  • 2. Concept Learning as Search:  Concept learning can be viewed as the task of searching through a large space of hypothesis implicitly defined by the hypothesis representation.  The goal of the concept learning search is to find the hypothesis that best fits the training examples.
  • 4. General-to-Specific Learning: Example Sky AirTemp Humidity Wind Water Forecast EnjoySport 1 Sunny Warm Normal Strong Warm Same Yes 2 Sunny Warm High Strong Warm Same Yes 3 Rainy Cold High Strong Warm Change No 4 Sunny Warm High Strong Cool Change Yes h1=(Sunny,?,?,Strong,?,?) h2=(Sunny,?,?,?,?,?) *h2 is more general than h1. *h2 imposes fewer constraints on the instance than h1.
  • 5. FIND-S: Finding a Maximally Specific Hypothesis: 1-Initialize h to the most specific hypotesis in h 2-for each positive training instance x - For each attribute constraint aj in h if the constraint aj is satisfied by x then do nothing else replace aj in h by the next more general constraint that is satisfied by x 3- output hypothesis h.
  • 6. Step 1: FIND-S: Example Sky AirTemp Humidity Wind Water Forecast EnjoySport 1 Sunny Warm Normal Strong Warm Same Yes 2 Sunny Warm High Strong Warm Same Yes 3 Rainy Cold High Strong Warm Change No 4 Sunny Warm High Strong Cool Change Yes Initialize h to the most specific hypotesis in h h0 = <Ø, Ø, Ø, Ø, Ø, Ø>
  • 8.
  • 9.
  • 10.  Version Space: The set of all valid hypotheses provided by an algorithm is called version space (VS)with respect to the hypothesis space Hand the given example set D.  Candidate-Elimination Algorithm: * The Candidate-Eliminationalgorithm finds all describable hypotheses that are consistent with the observed training examples. * Hypothesis is derived from examples regardless of whether x is positive or negative example
  • 12.  In principle, the LIST-THEN-ELIMINATE algorithm can be applied whenever the hypothesis space H is finite.  It is guaranteed to output all hypotheses consistent with the training data.  Unfortunately, it requires exhaustively enumerating all hypotheses in H-an unrealistic requirement for all but the most trivial hypothesis spaces.
  • 13. Candidate-Elimination Algorithm: •The CANDIDATE-ELIMINATION algorithm works on the same principleas the above LIST-THEN- ELIMINATE algorithm. •It employs a much more compact representation of the version space. •In this the version spaceis represented by its most general and least general members (Specific). •These members form general and specific boundary sets that delimit the version space within the partially ordered hypothesis space.
  • 14.
  • 15. Example : Example Sky AirTemp Humidity Wind Water Forecast EnjoySport 1 Sunny Warm Normal Strong Warm Same Yes 2 Sunny Warm High Strong Warm Same Yes 3 Rainy Cold High Strong Warm Change No 4 Sunny Warm High Strong Cool Change Yes
  • 16.
  • 17.
  • 18.
  • 19.
  • 20. What will Happen if the Training Contains errors ? Example Sky AirTemp Humidity Wind Water Forecast EnjoySport 1 Sunny Warm Normal Strong Warm Same Yes 2 Sunny Warm High Strong Warm Same No 3 Rainy Cold High Strong Warm Change No 4 Sunny Warm High Strong Cool Change Yes