WEKA Tutorial
- 1. Contributed by Yizhou Sun 2008 An Introduction to WEKA
- 2. Content What is WEKA? The Explorer: Preprocess data Classification Clustering Association Rules Attribute Selection Data Visualization References and Resources 04/26/10
- 3. What is WEKA? W aikato E nvironment for K nowledge A nalysis It’s a data mining/machine learning tool developed by Department of Computer Science, University of Waikato, New Zealand. Weka is also a bird found only on the islands of New Zealand. 04/26/10
- 4. Download and Install WEKA Website: http:// www.cs.waikato.ac.nz/~ml/weka/index.html Support multiple platforms (written in java): Windows, Mac OS X and Linux 04/26/10
- 5. Main Features 49 data preprocessing tools 76 classification/regression algorithms 8 clustering algorithms 3 algorithms for finding association rules 15 attribute/subset evaluators + 10 search algorithms for feature selection 04/26/10
- 6. Main GUI Three graphical user interfaces “ The Explorer” (exploratory data analysis) “ The Experimenter” (experimental environment) “ The KnowledgeFlow” (new process model inspired interface) 04/26/10
- 7. Content What is WEKA? The Explorer: Preprocess data Classification Clustering Association Rules Attribute Selection Data Visualization References and Resources 04/26/10
- 8. Explorer: pre-processing the data Data can be imported from a file in various formats: ARFF, CSV, C4.5, binary Data can also be read from a URL or from an SQL database (using JDBC) Pre-processing tools in WEKA are called “filters” WEKA contains filters for: Discretization, normalization, resampling, attribute selection, transforming and combining attributes, … 04/26/10
- 9. @relation heart-disease-simplified @attribute age numeric @attribute sex { female, male} @attribute chest_pain_type { typ_angina, asympt, non_anginal, atyp_angina} @attribute cholesterol numeric @attribute exercise_induced_angina { no, yes} @attribute class { present, not_present} @data 63,male,typ_angina,233,no,not_present 67,male,asympt,286,yes,present 67,male,asympt,229,yes,present 38,female,non_anginal,?,no,not_present ... WEKA only deals with “flat” files 04/26/10 Flat file in ARFF format
- 10. @relation heart-disease-simplified @attribute age numeric @attribute sex { female, male} @attribute chest_pain_type { typ_angina, asympt, non_anginal, atyp_angina} @attribute cholesterol numeric @attribute exercise_induced_angina { no, yes} @attribute class { present, not_present} @data 63,male,typ_angina,233,no,not_present 67,male,asympt,286,yes,present 67,male,asympt,229,yes,present 38,female,non_anginal,?,no,not_present ... WEKA only deals with “flat” files 04/26/10 numeric attribute nominal attribute
- 11. 04/26/10 University of Waikato
- 12. 04/26/10 University of Waikato
- 13. 04/26/10 University of Waikato
- 14. 04/26/10 University of Waikato
- 15. 04/26/10 University of Waikato
- 16. 04/26/10 University of Waikato
- 17. 04/26/10 University of Waikato
- 18. 04/26/10 University of Waikato
- 19. 04/26/10 University of Waikato
- 20. 04/26/10 University of Waikato
- 21. 04/26/10 University of Waikato
- 22. 04/26/10 University of Waikato
- 23. 04/26/10 University of Waikato
- 24. 04/26/10 University of Waikato
- 25. 04/26/10 University of Waikato
- 26. 04/26/10 University of Waikato
- 27. 04/26/10 University of Waikato
- 28. 04/26/10 University of Waikato
- 29. 04/26/10 University of Waikato
- 30. 04/26/10 University of Waikato
- 31. 04/26/10 University of Waikato
- 32. Explorer: building “classifiers” Classifiers in WEKA are models for predicting nominal or numeric quantities Implemented learning schemes include: Decision trees and lists, instance-based classifiers, support vector machines, multi-layer perceptrons, logistic regression, Bayes’ nets, … 04/26/10
- 33. April 26, 2010 This follows an example of Quinlan’s ID3 (Playing Tennis) Decision Tree Induction: Training Dataset
- 34. April 26, 2010 Output: A Decision Tree for “buys_computer” age? overcast student? credit rating? <=30 >40 no yes yes yes 31..40 no fair excellent yes no
- 35. Basic algorithm (a greedy algorithm) Tree is constructed in a top-down recursive divide-and-conquer manner At start, all the training examples are at the root Attributes are categorical (if continuous-valued, they are discretized in advance) Examples are partitioned recursively based on selected attributes Test attributes are selected on the basis of a heuristic or statistical measure (e.g., information gain ) April 26, 2010 Algorithm for Decision Tree Induction
- 36. 04/26/10 University of Waikato
- 37. 04/26/10 University of Waikato
- 38. 04/26/10 University of Waikato
- 39. 04/26/10 University of Waikato
- 40. 04/26/10 University of Waikato
- 41. 04/26/10 University of Waikato
- 42. 04/26/10 University of Waikato
- 43. 04/26/10 University of Waikato
- 44. 04/26/10 University of Waikato
- 45. 04/26/10 University of Waikato
- 46. 04/26/10 University of Waikato
- 47. 04/26/10 University of Waikato
- 48. 04/26/10 University of Waikato
- 49. 04/26/10 University of Waikato
- 50. 04/26/10 University of Waikato
- 51. 04/26/10 University of Waikato
- 52. 04/26/10 University of Waikato
- 53. 04/26/10 University of Waikato
- 54. 04/26/10 University of Waikato
- 55. 04/26/10 University of Waikato
- 56. 04/26/10 University of Waikato
- 57. 04/26/10 University of Waikato
- 58. Explorer: clustering data WEKA contains “clusterers” for finding groups of similar instances in a dataset Implemented schemes are: k -Means , EM, Cobweb, X -means, FarthestFirst Clusters can be visualized and compared to “true” clusters (if given) Evaluation based on loglikelihood if clustering scheme produces a probability distribution 04/26/10
- 59. Given k , the k-means algorithm is implemented in four steps: Partition objects into k nonempty subsets Compute seed points as the centroids of the clusters of the current partition (the centroid is the center, i.e., mean point , of the cluster) Assign each object to the cluster with the nearest seed point Go back to Step 2, stop when no more new assignment April 26, 2010 The K-Means Clustering Method
- 60. Demo Now. (Demo Online)
- 61. Explorer: finding associations WEKA contains an implementation of the Apriori algorithm for learning association rules Works only with discrete data Can identify statistical dependencies between groups of attributes: milk, butter bread, eggs (with confidence 0.9 and support 2000) Apriori can compute all rules that have a given minimum support and exceed a given confidence 04/26/10
- 62. Basic Concepts: Frequent Patterns itemset : A set of one or more items k-itemset X = {x 1 , …, x k } (absolute) support , or, support count of X: Frequency or occurrence of an itemset X (relative) support , s , is the fraction of transactions that contains X (i.e., the probability that a transaction contains X) An itemset X is frequent if X’s support is no less than a minsup threshold April 26, 2010 Customer buys diaper Customer buys both Customer buys beer Tid Items bought 10 Beer, Nuts, Diaper 20 Beer, Coffee, Diaper 30 Beer, Diaper, Eggs 40 Nuts, Eggs, Milk 50 Nuts, Coffee, Diaper, Eggs, Milk
- 63. Basic Concepts: Association Rules Find all the rules X Y with minimum support and confidence support , s , probability that a transaction contains X Y confidence , c, conditional probability that a transaction having X also contains Y Let minsup = 50%, minconf = 50% Freq. Pat.: Beer:3, Nuts:3, Diaper:4, Eggs:3, {Beer, Diaper}:3 April 26, 2010 Customer buys diaper Customer buys both Customer buys beer Nuts, Eggs, Milk 40 Nuts, Coffee, Diaper, Eggs, Milk 50 Beer, Diaper, Eggs 30 Beer, Coffee, Diaper 20 Beer, Nuts, Diaper 10 Items bought Tid Association rules: (many more!) Beer Diaper (60%, 100%) Diaper Beer (60%, 75%)
- 64. 04/26/10 University of Waikato
- 65. 04/26/10 University of Waikato
- 66. 04/26/10 University of Waikato
- 67. 04/26/10 University of Waikato
- 68. 04/26/10 University of Waikato
- 69. Explorer: attribute selection Panel that can be used to investigate which (subsets of) attributes are the most predictive ones Attribute selection methods contain two parts: A search method: best-first, forward selection, random, exhaustive, genetic algorithm, ranking An evaluation method: correlation-based, wrapper, information gain, chi-squared, … Very flexible: WEKA allows (almost) arbitrary combinations of these two 04/26/10
- 70. 04/26/10 University of Waikato
- 71. 04/26/10 University of Waikato
- 72. 04/26/10 University of Waikato
- 73. 04/26/10 University of Waikato
- 74. 04/26/10 University of Waikato
- 75. 04/26/10 University of Waikato
- 76. 04/26/10 University of Waikato
- 77. 04/26/10 University of Waikato
- 78. Explorer: data visualization Visualization very useful in practice: e.g. helps to determine difficulty of the learning problem WEKA can visualize single attributes (1-d) and pairs of attributes (2-d) To do: rotating 3-d visualizations (Xgobi-style) Color-coded class values “ Jitter” option to deal with nominal attributes (and to detect “hidden” data points) “ Zoom-in” function 04/26/10
- 79. 04/26/10 University of Waikato
- 80. 04/26/10 University of Waikato
- 81. 04/26/10 University of Waikato
- 82. 04/26/10 University of Waikato
- 83. 04/26/10 University of Waikato
- 84. 04/26/10 University of Waikato
- 85. 04/26/10 University of Waikato
- 86. 04/26/10 University of Waikato
- 87. 04/26/10 University of Waikato
- 88. 04/26/10 University of Waikato
- 89. References and Resources References: WEKA website: http://www.cs.waikato.ac.nz/~ml/weka/index.html WEKA Tutorial: Machine Learning with WEKA: A presentation demonstrating all graphical user interfaces (GUI) in Weka. A presentation which explains how to use Weka for exploratory data mining. WEKA Data Mining Book: Ian H. Witten and Eibe Frank, Data Mining: Practical Machine Learning Tools and Techniques (Second Edition) WEKA Wiki: http://weka.sourceforge.net/wiki/index.php/Main_Page Others: Jiawei Han and Micheline Kamber, Data Mining: Concepts and Techniques, 2nd ed.