CS 478 - Learning Rules PowerPoint Presentation, PPT - DocSlides

Download

  • Likes: 3
  • Views: 5
  • sherrill-nordquist

Description

CS 478 - Learning Rules PowerPoint Presentation, PPT - DocSlides - 1. Learning Sets of Rules. CS 478 - Learning Rules. 2. Learning Rules. If (Color = Red) and (Shape = round) then Class is A. If (Color = Blue) and (Size = large) then Class is B. If (Shape = square) then Class is A.

Download this presentation

CS 478 - Learning Rules PowerPoint Presentation, PPT - DocSlides

Click below link (As may be) to download this presentation.

Download Note - The PPT/PDF document "CS 478 - Learning Rules PowerPoint Prese..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.

CS 478 - Learning Rules PowerPoint Presentation, PPT - DocSlides

    Slide1

    CS 478 - Learning Rules

    1

    Learning Sets of Rules

    Slide2

    CS 478 - Learning Rules

    2

    Learning Rules

    If (Color = Red) and (Shape = round) then Class is A

    If (Color = Blue) and (Size = large) then Class is B

    If (Shape = square) then Class is A

    Natural and intuitive hypotheses

    Comprehensibility - Easy to understand?

    Slide3

    CS 478 - Learning Rules

    3

    Learning Rules

    If (Color = Red) and (Shape = round) then Class is A

    If (Color = Blue) and (Size = large) then Class is B

    If (Shape = square) then Class is A

    Natural and intuitive hypotheses

    Comprehensibility - Easy to understand?

    Ordered (Prioritized) rules - default at the bottom, common but not so easy to comprehend

    Unordered rules

    Theoretically easier to understand, except must

    Force consistency, or

    Create a separate unordered list for each output class and use a tie-break scheme when multiple lists are matched

    Slide4

    CS 478 - Learning Rules

    4

    Sequential Covering Algorithms

    There are a number of rule learning algorithms based on different variations of sequential covering

    CN2, AQx, etc.

    Find a “good” rule for the current training set

    Delete covered instances (or those covered correctly) from the training set

    Go back to 1 until the training set is empty or until no more “good” rules can be found

    Slide5

    CS 478 - Learning Rules

    5

    Finding “Good” Rules

    The large majority of instances covered by the rule infer the same output class

    Rule covers as many instances as possible (general vs specific rules)

    Rule covers enough instances (statistically significant)

    Example rules and approaches?

    How to find good rules efficiently? - General to specific search is common

    Continuous features - some type of ranges/discretization

    Slide6

    CS 478 - Learning Rules

    6

    Common Rule “Goodness” Approaches

    Relative frequency: nc/nm-estimate of accuracy (better when n is small): where p is the prior probability of a random instance having the output class of the proposed rule, penalizes rules with small n, Laplacian common: (nc+1)/(n+|C|) (i.e. m = 1/pc)Entropy - Favors rules which cover a large number of examples from a single class, and few from othersEntropy can be better than relative frequencyImproves consequent rule induction. R1:(.7,.1,.1,.1) R2 (.7,0.3,0) - entropy selects R2 which makes for better subsequent specializations during later rule growthEmpirically, rules of low entropy have higher significance than relative frequency, but Laplacian often better than entropy

    Slide7

    CS 478 - Learning Rules

    7

    Slide8

    CS 478 - Learning Rules

    8

    Slide9

    CS 478 - Learning Rules

    9

    Slide10

    CS 478 - Learning Rules

    10

    Learning First Order Rules

    Inductive Logic Programming

    Propositional vs. first order rules

    1st order allows variables in rules

    If Color of object1 =

    x

    and Color of object 2 =

    x

    then Class is A

    More expressive

    FOIL - Uses a sequential covering approach from general to specific which allows the addition of literals with variables

    Slide11

    CS 478 - Learning Rules

    11

    RIPPER

    Slide12

    CS 478 - Learning Rules

    12

    Slide13

    Insert Rules at Top or Bottom

    Typically would like focused exception rules higher and more general rules lower in the listTypical (CN2): Delete all instances covered by a rule during learningPutting new rule on the bottom (i.e. early learned rules stay on top) makes sense since this rule is rated good only after removing all instances covered by previous rules, (i.e. instances which can get by the earlier rules).Still should get exceptions up top and general rules lower in the list because exceptions will achieve a higher score and thus be added first (assuming statistical significance) than a general rule which has to cover more cases. Even though E keeps getting diminished there should still be enough data to support reasonable general rules later (in fact the general rules should get increasing scores after true exceptions are removed).Highest scoring rules: Somewhat specific, high accuracy, sufficient coverageMedium scoring rules: General and specific with reasonable accuracy and coverageLow scoring rules: Specific with low coverage, and general with low accuracy

    13

    Slide14

    Rule Order - Continued

    If delete only correct instances covered by a rulePutting new rule on the the top (i.e. first learned rule stays on bottom) could make sense because we could learn exception rules for those instances not covered by general rules at the bottomThis only works if the rule placed at the bottom is truly more general than the later rules (i.e. many novel instances will slide past the more exceptional rules and get covered by the general rules at the bottom)Sort after: (Mitchell) Proceed with care because rules were learned based on specific subsets of the training setOther variations possible, but many could be problematic because there are an exponential number of possible orderingsAlso can do unordered lists with tie-breaking mechanisms

    CS 478 - Learning Rules

    14

Report this Document.

  • Creating Content

    Creating Content

    733-views
  • Clipper: A Low Latency Online Prediction Serving System

    Clipper: A Low Latency Online Prediction Serving System

    703-views
  • Claims, Evidence, and Warrants

    Claims, Evidence, and Warrants

    612-views
  • You Can Choose Engineered Timber Flooring In Melbourne

    You Can Choose Engineered Timber Flooring In Melbourne

    790-views
  • Gaucho Round-Up

    Gaucho Round-Up

    742-views
  • SCADA in electrical power delivery

    SCADA in electrical power delivery

    763-views
  • A Century of Turmoil

    A Century of Turmoil

    735-views
  • Theories of evolution

    Theories of evolution

    795-views
  • Drones Collecting Cell Phone Data in

    Drones Collecting Cell Phone Data in

    787-views
  • Fossil Collecting:

    Fossil Collecting:

    751-views