An Introduction to

WEKA
Introduction to Machine Learning

Dr. Hikmat Ullah Khan

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 1 10/11/2018

 Content
O What is WEKA?
O The Explorer:
O Preprocess data
O Classification
O Clustering
O Association Rules
O Attribute Selection
O Data Visualization
O References and Resources
O ?

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 2

 What is WEKA?
O Waikato Environment for Knowledge
Analysis
O It’s a data mining/machine learning tool
O Developed by Department of Computer
Science, University of Waikato, New Zealand.
O Weka is also a bird found only on the islands
of New Zealand.

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 3

 WEKA only deals with “flat” files
@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
...Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 4
 WEKA only deals with “flat” files
@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
...
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 5
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 6
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 7
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 8
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 9
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 10
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 11
 Explorer: building “classifiers”
O Classifiers in WEKA are models for predicting
nominal or numeric quantities
O Implemented learning schemes include:
O Decision trees and lists, instance-based classifiers,
support vector machines, multi-layer perceptrons,
logistic regression, Bayes’ nets, …
O Latest classifiers include:
O Bagging, boosting, stacking, error-correcting output
codes, locally weighted learning, …
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 12
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 13
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 14
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 15
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 16
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 17
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 18
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 19
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 20
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 21
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 22
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 23
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 24
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 25
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 26
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 27
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 28
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 29
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 30
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 31
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 32
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 33
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 34
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 35
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 36
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 37
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 38
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 39
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 40
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 41
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 42
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 43
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 44
 Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 45

QuickTim e™ an d a TIFF (LZW) de com pressor a re ne eded to see t his picture.

 Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 46

QuickTim e™ an d a TIFF (LZW) de com pressor a re ne eded to see t his picture.

 Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 47

QuickTim e™ an d a TIFF (LZW) de com pressor a re ne eded to see t his picture.

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 48
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 49
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 50
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 51
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 52
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 53
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 54
Quic kTime™and a TIFF(LZW) decom pres sor are needed to s ee this pic tu re .

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 55

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 56
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 57
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 58
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 59
QuickTim e™ an d a TIFF (LZW) decom pressor are nee ded t o see th is pict ure.

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 60

QuickTim e™ an d a TIFF (LZW) decom pressor are nee ded t o see th is pict ure.

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 61

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 62
QuickTim e™ an d a TIFF (LZW) decom pressor are nee ded t o see th is pict ure.

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 63

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 64
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 65
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 66
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 67
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 68
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 69
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 70
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 71
 Main Features
O 49 data preprocessing tools
O 76 classification/regression
algorithms
O 8 clustering algorithms
O 3 algorithms for finding association
rules
O 15 attribute/subset evaluators + 10
search algorithms for feature
selection
Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 72
 Thanks

Dr. Hikmat Ullah Khan ( Thanks :University of Waikato) 10/11/2018 73