Download presentation
Presentation is loading. Please wait.
1
Department of Computer Science, University of Waikato, New Zealand Eibe Frank WEKA: A Machine Learning Toolkit The Explorer Classification and Regression Clustering Association Rules Attribute Selection Data Visualization The Experimenter The Knowledge Flow GUI Conclusions Machine Learning with WEKA
2
6/9/2015University of Waikato2 WEKA: the bird Copyright: Martin Kramer (mkramer@wxs.nl)
3
6/9/2015University of Waikato3 WEKA: the software Machine learning/data mining software written in Java (distributed under the GNU Public License) Used for research, education, and applications Complements “Data Mining” by Witten & Frank Main features: Comprehensive set of data pre-processing tools, learning algorithms and evaluation methods Graphical user interfaces (incl. data visualization) Environment for comparing learning algorithms
4
6/9/2015University of Waikato4 WEKA: versions There are several versions of WEKA: WEKA 3.0: “book version” compatible with description in data mining book WEKA 3.2: “GUI version” adds graphical user interfaces (book version is command-line only) WEKA 3.3: “development version” with lots of improvements This talk is based on the latest snapshot of WEKA 3.3 (soon to be WEKA 3.4)
5
6/9/2015University of Waikato5 @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
6
6/9/2015University of Waikato6 @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
7
6/9/2015University of Waikato7
8
6/9/2015University of Waikato8
9
6/9/2015University of Waikato9
10
6/9/2015University of Waikato10 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, …
11
6/9/2015University of Waikato11
12
6/9/2015University of Waikato12
13
6/9/2015University of Waikato13
14
6/9/2015University of Waikato14
15
6/9/2015University of Waikato15
16
6/9/2015University of Waikato16
17
6/9/2015University of Waikato17
18
6/9/2015University of Waikato18
19
6/9/2015University of Waikato19
20
6/9/2015University of Waikato20
21
6/9/2015University of Waikato21
22
6/9/2015University of Waikato22
23
6/9/2015University of Waikato23
24
6/9/2015University of Waikato24
25
6/9/2015University of Waikato25
26
6/9/2015University of Waikato26
27
6/9/2015University of Waikato27
28
6/9/2015University of Waikato28
29
6/9/2015University of Waikato29
30
6/9/2015University of Waikato30
31
6/9/2015University of Waikato31
32
6/9/2015University of Waikato32 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, … “Meta”-classifiers include: Bagging, boosting, stacking, error-correcting output codes, locally weighted learning, …
33
6/9/2015University of Waikato33
34
6/9/2015University of Waikato34
35
6/9/2015University of Waikato35
36
6/9/2015University of Waikato36
37
6/9/2015University of Waikato37
38
6/9/2015University of Waikato38
39
6/9/2015University of Waikato39
40
6/9/2015University of Waikato40
41
6/9/2015University of Waikato41
42
6/9/2015University of Waikato42
43
6/9/2015University of Waikato43
44
6/9/2015University of Waikato44
45
6/9/2015University of Waikato45
46
6/9/2015University of Waikato46
47
6/9/2015University of Waikato47
48
6/9/2015University of Waikato48
49
6/9/2015University of Waikato49
50
6/9/2015University of Waikato50
51
6/9/2015University of Waikato51
52
6/9/2015University of Waikato52
53
6/9/2015University of Waikato53
54
6/9/2015University of Waikato54
55
6/9/2015University of Waikato55
56
6/9/2015University of Waikato56
57
6/9/2015University of Waikato57
58
6/9/2015University of Waikato58
59
6/9/2015University of Waikato59
60
6/9/2015University of Waikato60
61
6/9/2015University of Waikato61
62
6/9/2015University of Waikato62
63
6/9/2015University of Waikato63
64
6/9/2015University of Waikato64
65
6/9/2015University of Waikato65
66
6/9/2015University of Waikato66
67
6/9/2015University of Waikato67
68
6/9/2015University of Waikato68
69
6/9/2015University of Waikato69
70
6/9/2015University of Waikato70
71
6/9/2015University of Waikato71
72
6/9/2015University of Waikato72
73
6/9/2015University of Waikato73
74
6/9/2015University of Waikato74
75
6/9/2015University of Waikato75
76
6/9/2015University of Waikato76
77
6/9/2015University of Waikato77
78
6/9/2015University of Waikato78
79
6/9/2015University of Waikato79
80
6/9/2015University of Waikato80
81
6/9/2015University of Waikato81
82
6/9/2015University of Waikato82
83
6/9/2015University of Waikato83
84
6/9/2015University of Waikato84
85
6/9/2015University of Waikato85
86
6/9/2015University of Waikato86
87
6/9/2015University of Waikato87
88
6/9/2015University of Waikato88
89
6/9/2015University of Waikato89
90
6/9/2015University of Waikato90
91
6/9/2015University of Waikato91
92
6/9/2015University of Waikato92 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
93
6/9/2015University of Waikato93
94
6/9/2015University of Waikato94
95
6/9/2015University of Waikato95
96
6/9/2015University of Waikato96
97
6/9/2015University of Waikato97
98
6/9/2015University of Waikato98
99
6/9/2015University of Waikato99
100
6/9/2015University of Waikato100
101
6/9/2015University of Waikato101
102
6/9/2015University of Waikato102
103
6/9/2015University of Waikato103
104
6/9/2015University of Waikato104
105
6/9/2015University of Waikato105
106
6/9/2015University of Waikato106
107
6/9/2015University of Waikato107
108
6/9/2015University of Waikato108 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
109
6/9/2015University of Waikato109
110
6/9/2015University of Waikato110
111
6/9/2015University of Waikato111
112
6/9/2015University of Waikato112
113
6/9/2015University of Waikato113
114
6/9/2015University of Waikato114
115
6/9/2015University of Waikato115
116
6/9/2015University of Waikato116 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
117
6/9/2015University of Waikato117
118
6/9/2015University of Waikato118
119
6/9/2015University of Waikato119
120
6/9/2015University of Waikato120
121
6/9/2015University of Waikato121
122
6/9/2015University of Waikato122
123
6/9/2015University of Waikato123
124
6/9/2015University of Waikato124
125
6/9/2015University of Waikato125 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
126
6/9/2015University of Waikato126
127
6/9/2015University of Waikato127
128
6/9/2015University of Waikato128
129
6/9/2015University of Waikato129
130
6/9/2015University of Waikato130
131
6/9/2015University of Waikato131
132
6/9/2015University of Waikato132
133
6/9/2015University of Waikato133
134
6/9/2015University of Waikato134
135
6/9/2015University of Waikato135
136
6/9/2015University of Waikato136
137
6/9/2015University of Waikato137
138
6/9/2015University of Waikato138 Performing experiments Experimenter makes it easy to compare the performance of different learning schemes For classification and regression problems Results can be written into file or database Evaluation options: cross-validation, learning curve, hold-out Can also iterate over different parameter settings Significance-testing built in!
139
6/9/2015University of Waikato139
140
6/9/2015University of Waikato140
141
6/9/2015University of Waikato141
142
6/9/2015University of Waikato142
143
6/9/2015University of Waikato143
144
6/9/2015University of Waikato144
145
6/9/2015University of Waikato145
146
6/9/2015University of Waikato146
147
6/9/2015University of Waikato147
148
6/9/2015University of Waikato148
149
6/9/2015University of Waikato149
150
6/9/2015University of Waikato150
151
6/9/2015University of Waikato151
152
6/9/2015University of Waikato152 The Knowledge Flow GUI New graphical user interface for WEKA Java-Beans-based interface for setting up and running machine learning experiments Data sources, classifiers, etc. are beans and can be connected graphically Data “flows” through components: e.g., “data source” -> “filter” -> “classifier” -> “evaluator” Layouts can be saved and loaded again later
153
6/9/2015University of Waikato153
154
6/9/2015University of Waikato154
155
6/9/2015University of Waikato155
156
6/9/2015University of Waikato156
157
6/9/2015University of Waikato157
158
6/9/2015University of Waikato158
159
6/9/2015University of Waikato159
160
6/9/2015University of Waikato160
161
6/9/2015University of Waikato161
162
6/9/2015University of Waikato162
163
6/9/2015University of Waikato163
164
6/9/2015University of Waikato164
165
6/9/2015University of Waikato165
166
6/9/2015University of Waikato166
167
6/9/2015University of Waikato167
168
6/9/2015University of Waikato168
169
6/9/2015University of Waikato169
170
6/9/2015University of Waikato170
171
6/9/2015University of Waikato171
172
6/9/2015University of Waikato172
173
6/9/2015University of Waikato173 Conclusion: try it yourself! WEKA is available at http://www.cs.waikato.ac.nz/ml/weka Also has a list of projects based on WEKA WEKA contributors: Abdelaziz Mahoui, Alexander K. Seewald, Ashraf M. Kibriya, Bernhard Pfahringer, Brent Martin, Peter Flach, Eibe Frank,Gabi Schmidberger,Ian H. Witten, J. Lindgren, Janice Boughton, Jason Wells, Len Trigg, Lucio de Souza Coelho, Malcolm Ware, Mark Hall,Remco Bouckaert, Richard Kirkby, Shane Butler, Shane Legg, Stuart Inglis, Sylvain Roy, Tony Voyle, Xin Xu, Yong Wang, Zhihai Wang
Similar presentations
