**INTRODUCTION:**

1.Well-posed learning problems

2.Designing a learning system

3.Choosing the training experience

4.Choosing the target function

5.choosing a representation for the target function

6.Choosing a function approximation algorithm

7.The final design

8.Perspectives and isssues in machine learning

9.Issues in machine learning

10.How to read this book

11.Summary and further reading

12.Exercises

13.Reference

**CONCEPT LEARNING AND THE GENERAL-TO-SPECIFIC ORDERING :**

1.Introduction

2.A concept learning task

3.Notation

4.The inductive learning hypothesis

5.Concept learning as search

6.General-specific ordering to hypotheses

7.Finding a maximally specific hypothesis

8.Version spaces and the candidate -elimination algorithm

9.Representation

10.The list- then-eliminate algorithm

11.A more compact representation for version spaces

12.Candidate – elimination learning algorithm

13.An illustrative example

14.Remarks on version spaces and candidate – elimination

15.Will the candidate – elemination algorithm converge to the correct hypothesis?

16.What training example should the learner request next?

17.How can partially learned concept be used?

18.Inductive bias

19.A biased hypothesis space

20.An unbiased learner

21.The futility of bias – free learning

22.Summary and further reading

23.Exercise

24.References

**DECISION TREE LEARNING:**

1.Introduction

2.Decision tree representation

3.Appropriate problems for decision tree learning

4.The basic decision tree learning algorithm

5.Which attribute is the best classifier ?

6.An illustrate example

7.Hypothesis space search in decision tree learning

8.Inductive bias in decision tree learning

9.Restriction biases and preference biases

10.Why prefer short hypotheses?

11.Issues in decision tree learning

12.Avoiding overfitting the data

13.Incorporating continuous -valued attributes

14.Alternative measures for selecting attributes

15.Handling training examples with missing attribute values

16.Handling attributes with differing costs

17.Summary and further reading

18.Exercise

19.Reference

**ARTIFICIAL NEURAL NETWORKS:**

1.Introduction

2.Biological motivation

3.Neural network representations

4.Appropriate problems for neural network learning

5.Perceptrons

6.Representational power of perceptrons

7.The perceptron training rule

8.Gradient descent and the delta rule

9.Remarks

10.Multilayer networks and the backpropagation algorithm

11.A differentiable threshold unit

12.The backpropagation algorithm

13.Derivation of the backpropagation rule

14.Remarks on the backpropagation algorithm

15.Convergence and local minima

16.Representational power of feedforwad networks

17.Hypothesis space search and inductive bias

18.Hidden layer representation

19.Generalization ,overfitting,and stopping criterion

20.An illustrative example :face recognition

21.The task

22.Design choices

23.Learned hidden representations

24.Advanced topics in artificial neural networks

25.Alternative error functions

26.Recurrent networks

27.Dynamically modifying network structure

28.Summary and further reading

29.Exercise

30.References

**EVALUATING HYPOTHESES **

1.Motivation

2.Estimating hypothesis accuracy

3.Sample error and true error

4.Confidence intervals for discrete-valued hypotheses

5.Basics of sampling theory

6.Error estimation and estimating binomial proportions

7.The binomial distribution

8.Mean and variance

9.Estimators,bias,and variance

10.Confidence intervals

11.Two-sided and one -sided bounds

12.A general approach for deriving confidence intervals

13.Central limit theorem

14.Difference in error of two hypotheses

15.Hypothesis testing

16.Comparing learning algorithms

17.Paired t tests

18.Practical considerations

19.Summary and further reading

20.Exercises

21.References

**BAYESIAN LEARNING:**

1.Introduction

2.Bayes theorem

3.An example

4.Bayes theoram and concept learning

5.Brute-force bayes concept learning

6.MAP hypotheses and consistent learners

7.Maximum likelihood and least- squared error hypotheses

8.Maximum likelihood hypotheses for predicting probablities

9.Gradient search to maximize likelihood in neural net

10.minimum description length principal

11.Bayes optimal classifier

12.Gibbs algorithm

13.Naive bayes classifier

14.An illustrative example

15.An example :learning to classify textv

16.Experimental results

17.Bayesian belief networks

18.Conditional independence

19.Representation

20.Inference

21.Learning bayesian belief networks

22.Gradient ascent training of bayesian networks

23.Learning the structure of bayesian networks

24.The EM algorithm

25.Estimating means of k gaussians

26.General statement of EM algorithm

27.Derivation of the k means algorithm

28.Summary and further reading

29.Exercises

30.References

**COMPUTATIONAL LEARNING THEORY :**

1.Introduction

2.Probably learning an approximately correct hypothesis

3.The problem setting

4.Error of a hypothesis

5.PAC learnability

6.Sample complexity for finite hypothesis spaces

7.Agnostic learning and inconsistent hypotheses

8.Conjunctions of boolean literals are PAC =learnable

9.PAC -learnabilty of other concept classes

10.Sample complexity for infinite hypothesis spaces

11.Shattering a set of instances

12.The vapnik-chervonenkis dimension

13.Sample complexity and the VC dimension

14.VC dimension or neural networks

15.The mistake bound model for learning

16.Mistake bound for the FIND-S algorithm

17.Mistake bound for the HALVING algorithm

18.Optimal mistake bounds

19.WEIGHTED _MAJORITY algorithm

20.Summary and further reading

21.Exercise

22.References

**INSTANCE -BASED LEARNING:**

1.Introduction

2.K-nearest neighbor learning

3.Distance-weighted nearest neighbour algorithm

4.Remarks onn k-nearest neighbor algorithm

5.A note on terminology

6.Locally weighted regression

7.Locally weighted linear regression

8.Remarks on locally weighted regression

9.Redial basics functions

10.Case-based reasoning

11.Remarks on lazy and eager learning

12.Summary and further reading

13.Exercise

14.References

**GENETIC ALGORITHMS :**

1.Motivation

2.Genetic algorithms

3.Representing hypotheses

4.Genetic operators

5.Fitness function and selection

6.An llustrative example

7.Extensions

8.Hypothesis space search

9.Population evolution and the schema theorem

10.Genetic programming

11.Representing programs

12.Illustrative example

13.Remarks on genetic programming

14.Models of evolution learning

15.Lamarckian evolution

16.Baldwin effect

17.Parellelizing genetic algorithms

18.Summary and further reading

19.Exercises

20.References

**LEARNING SETS OF RULES:**

1.Introduction

2.Sequential covering algorithms

3.General to specific beam search

4.Variations

5.Learning rules sets:summary

6.Learning first-order rules

7.First-order learn clauses

8.Terminology

9.Learning sets of first-order rules:FOIL

10.Generating candidate specialization in foil

11.Guiding the search in foil

12.Learning recursive rule sets

13.Summary of FOIL

14.Introduction as inverted deduction

15.First-order resolution

16.Inverting resolution:first-order case

17.Summary of inverse resolution

18.Generalization ,subsumption ,andentailment

19.Progol

20.Summary and further reading

21.Exercise

22.References

**ANALYTICAL LEARNING :**

1.Introduction

2.Inductive and analytical learning problems

3.Learning with perfect domain theories :prolog -EBG

4.An illustrative trace

5.Remarks on explanation-based learning

6.Discovering new features

7.Deductive learning

8.Inductive bias in explanation -based learning

9.Knowledge level learning

10.Explanation -based learning of search control knowledge

11.Summary and further reading

12.Exercises

13.References

**COMBINING INDUCTIVE AND ANALLYTICAL LEARNING:**

1.Motivation

2.Inductive-analytical approaches to learning

3.The learning problem

4.Hyphothesis space search

5.Using prior knowledge to initialize the hypothesis

6.The KBANN algorithm

7.An illustrative example

8.Remarks

9.Using prior knowledge to alter the search objective

10.The TANGENTPROP algorithm

11.An illustrative example

12.The EBNN algorithm

13.Remarks

14.Using prior knowledge ton augment search operators

15. The FOCL algorithm

16.Remarks

17.State of the art

18.Summary and further reading

19.Exercises

20.References

**REIGNFORCEMENT LERANING:**

1.Introduction

2.the learning task

3.Q learning

4.The Q function

5.An algorithm for learning Q

6.An illustrative example

7.Convergence

8.Experimentation strategies

9.Updating sequence

10.Nondeterministic rewards and actions

11.Temporal difference learning

12.Generalizing from examples

13.Relationship to dynamic programming

14.Summary and further reading

15.Exercise

16.Reference