Atnaujinkite slapukų nuostatas

Introduction to Learning Classifier Systems 2017 ed. [Minkštas viršelis]

4.00/5 (3 ratings by Goodreads)
,
  • Formatas: Paperback / softback, 123 pages, aukštis x plotis: 235x155 mm, weight: 2234 g, 4 Illustrations, color; 23 Illustrations, black and white; XIII, 123 p. 27 illus., 4 illus. in color., 1 Paperback / softback
  • Serija: SpringerBriefs in Intelligent Systems
  • Išleidimo metai: 06-Sep-2017
  • Leidėjas: Springer-Verlag Berlin and Heidelberg GmbH & Co. K
  • ISBN-10: 3662550067
  • ISBN-13: 9783662550069
Kitos knygos pagal šią temą:
Introduction to Learning Classifier Systems 2017 ed.Didesnis paveikslėlis
  • Formatas: Paperback / softback, 123 pages, aukštis x plotis: 235x155 mm, weight: 2234 g, 4 Illustrations, color; 23 Illustrations, black and white; XIII, 123 p. 27 illus., 4 illus. in color., 1 Paperback / softback
  • Serija: SpringerBriefs in Intelligent Systems
  • Išleidimo metai: 06-Sep-2017
  • Leidėjas: Springer-Verlag Berlin and Heidelberg GmbH & Co. K
  • ISBN-10: 3662550067
  • ISBN-13: 9783662550069
Kitos knygos pagal šią temą:
Pastovi nuoroda: https://www.kriso.lt/db/9783662550069.html
Raktažodžiai:
This accessible introduction shows the reader how to understand, implement, adapt, and apply Learning Classifier Systems (LCSs) to interesting and difficult problems. The text builds an understanding from basic ideas and concepts. The authors first explore learning through environment interaction, and then walk through the components of LCS that form this rule-based evolutionary algorithm. The applicability and adaptability of these methods is highlighted by providing descriptions of common methodological alternatives for different components that are suited to different types of problems from data mining to autonomous robotics. 





The authors have also paired exercises and a simple educational LCS (eLCS) algorithm (implemented in Python) with this book. It is suitable for courses or self-study by advanced undergraduate and postgraduate students in subjects such as Computer Science, Engineering, Bioinformatics, and Cybernetics, and by researchers, data analysts, andmachine learning practitioners.

Recenzijos

Introduction to Learning Classifier Systems is an excellent textbook and introduction to Learning Classifier Systems. The book is completed with Python code available through a link included in the book. Urbanowicz and Browne recommend their book for undergraduate and postgraduate students, data analysts, and machine learning practitioners alike. (Analķa Amandi, Genetic Programming and Evolvable Machines, Vol. 19 (4), December, 2018)

1 LCSs in a Nutshell
1(20)
1.1 A Non-trivial Example Problem: The Multiplexer
1(3)
1.2 Key Elements
4(3)
1.2.1 Environment
4(1)
1.2.2 Rules, Matching, and Classifiers
4(2)
1.2.3 Discovery Component - Evolutionary Computation
6(1)
1.2.4 Learning Component
7(1)
1.3 LCS Functional Cycle
7(3)
1.4 Post-training
10(3)
1.4.1 Rule Compaction
10(1)
1.4.2 Prediction
11(1)
1.4.3 Evaluation
11(1)
1.4.4 Interpretation
12(1)
1.5 Code Exercises (eLCS)
13(8)
2 LCS Concepts
21(20)
2.1 Learning
22(2)
2.1.1 Modeling with a Ruleset
23(1)
2.2 Classifier
24(6)
2.2.1 Rules
24(3)
2.2.2 Representation and Alphabet
27(1)
2.2.3 Generalisation
28(2)
2.3 System
30(5)
2.3.1 Interaction with Problems
30(3)
2.3.2 Cooperation of Classifiers
33(1)
2.3.3 Competition Between Classifiers
34(1)
2.4 Problem Properties
35(4)
2.4.1 Problem Complexity
35(2)
2.4.2 Applications Overview
37(2)
2.5 Advantages
39(1)
2.6 Disadvantages
40(1)
3 Functional Cycle Components
41(30)
3.1 Evolutionary Computation and LCSs
42(2)
3.2 Initial Considerations
44(1)
3.3 Basic Alphabets for Rule Representation
45(6)
3.3.1 Encoding for Binary Alphabets
45(3)
3.3.2 Interval-Based
48(3)
3.4 Matching
51(1)
3.5 Covering
52(1)
3.6 Form a Correct Set or Select an Action
53(4)
3.6.1 Explore vs. Exploit
54(2)
3.6.2 Action Selection
56(1)
3.7 Performing the Action
57(1)
3.8 Update
58(1)
3.8.1 Numerosity of Rules
58(1)
3.8.2 Fitness Sharing
59(1)
3.9 Selection for Rule Discovery
59(3)
3.9.1 Parent Selection Methods
60(2)
3.10 Rule Discovery
62(6)
3.10.1 When to Invoke Rule Discovery
63(1)
3.10.2 Identifying Building Blocks of Knowledge
64(1)
3.10.3 Mutation
65(1)
3.10.4 Crossover
66(1)
3.10.5 Initialising Offspring Classifiers
67(1)
3.10.6 Other Rule Discovery
68(1)
3.11 Subsumption
68(1)
3.12 Deletion
69(1)
3.13 Summary
70(1)
4 LCS Adaptability
71(32)
4.1 LCS Pressures
71(3)
4.2 Michigan-Style vs. Pittsburgh-Style LCSs
74(2)
4.3 Michigan-Style Approaches
76(11)
4.3.1 Michigan-Style Supervised Learning (UCS)
76(2)
4.3.2 Updates with Time-Weighted Recency Averages
78(1)
4.3.3 Michigan-Style Reinforcement Learning (e.g. XCS)
79(8)
4.4 Pittsburgh-Style Approaches
87(2)
4.4.1 GAssist and BioHEL
87(1)
4.4.2 GABIL, GALE, and A-PLUS
88(1)
4.5 Strength- vs. Accuracy-Based Fitness
89(1)
4.5.1 Strength-Based
89(1)
4.5.2 Accuracy-Based
90(1)
4.6 Niche-Based Rule Discovery
90(2)
4.7 Single- vs. Multi-step Learning
92(6)
4.7.1 Sense, Plan, Act
93(2)
4.7.2 Delayed Reward
95(2)
4.7.3 Anticipatory Classifier Systems
97(1)
4.8 Computed Alphabets
98(3)
4.8.1 S-Expression and Genetic Programming
98(1)
4.8.2 Artificial Neural Networks
99(1)
4.8.3 Computed Prediction
99(1)
4.8.4 Computed Action
100(1)
4.8.5 Code Fragments
100(1)
4.9 Environment Considerations
101(2)
5 Applying LCSs
103
5.1 LCS Setup
104(10)
5.1.1 Run Parameter `Sweet Spots'
110(3)
5.1.2 Hybridise or Die
113(1)
5.2 Tuning
114(1)
5.3 Troubleshooting
115(2)
5.3.1 Lack of Convergence
116(1)
5.4 Where to Now?
117(6)
5.4.1 Workshops and Conferences
117(1)
5.4.2 Books, Journals, and Select Reviews
118(3)
5.4.3 Websites and Software
121(1)
5.4.4 Collaborate
122(1)
5.5 Concluding Remarks
123
Ryan Urbanowicz is a postdoctoral research associate in the Dept. of Biostatistics, Epidemiology, and Informatics in the Perelman School of Medicine at the University of Pennsylvania. He received his PhD in Genetics from Dartmouth College, and a B.S. and M.Eng. in Biological Engineering from Cornell University. His areas of research include bioinformatics, data mining, machine learning, evolutionary algorithms, learning classifier systems, data visualization, and epidemiology. He has cochaired the Intl. Workshop on Learning Classifier Systems and presented LCS tutorials at GECCO.

Will Browne is an Associate Professor in the School of Engineering and Computer Science of Victoria University of Wellington. He received his Eng.D. from Cardiff University. His main area of research is applied cognitive systems, in particular cognitive robotics, Learning Classifier Systems (LCSs), and modern heuristics for industrial application. He has cochaired the Intl. Workshop on Learning Classifier Systems, and chaired the Genetics-Based Machine Learning track and copresented the LCS tutorial at GECCO.