Machine Learning with WEKA An Introduction

2005/9/30

WEKA: the bird

Copyright: Martin Kramer ([email protected])

WEKA is available at
http://www.cs.waikato.ac.nz/ml/weka/

The format of Dataset in WEKA(1)

@relation heart-disease-simplified @attribute @attribute @attribute @attribute @attribute @attribute age numeric sex { female, male} chest_pain_type { typ_angina, asympt, non_anginal, atyp_angina} cholesterol numeric exercise_induced_angina { no, yes} 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 ...

The format of Dataset in WEKA(2)

@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 ...

Data can be imported from a file in various formats: ARFF, CSV, C4.5, binary

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,

Pre-processing tools in WEKA are called filters: including discretization, normalization, resampling, attribute selection, transforming and combining attributes,

Visualize class distribution for each attribute

Click here to choose filter algorithm

Click here to set the parameter for filter algorithm

Set parameter

apply the filter algorithm

Equal frequence

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,

Choose classification algorithm

J48 : Decision tree algorithm

Click here to set parameter for decision tree

Information about decision tree algorithm

Output setting

Start to build classifier

Fitted result

Click right bottom to view more information

View fitted tree

Crosses : correctly classified instances Squares : incorrectly classified instances

SMO : Support Vector Machine algorithm

Choose RBF kernel

Click right bottom to view more information

clustering data
WEKA contains clusterers for finding groups of similar instances in a dataset Implemented schemes are:
FarthestFirst

k-Means, EM, Cobweb, X-means,

Clusters can be visualized and compared to true clusters (if given) Evaluation based on loglikelihood if clustering scheme produces a probability distribution

clustering data

Choose clustering algorithm

Click here to set parameter

Start to perform clustering

Click right bottom to view more information

View cluster assignments

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

Finding associations

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

Choose attribute evalutor

Choose search method

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

Adjust view point size

Click here to view single result

Choose range to magnify local result

submit to magnify local result

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!

Click here to run experiment

Step 1 : Set output file Step 3 : choose algorithm Step 2 : add dataset

Experiment status

View the experiment result

Click here to perform test

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

Step 1 : choose data source

Step 2 : choose data source format

Step 3 : set data connection

Visualize data

Step 4 : choose evaluation

Step 5 : choose filter method

Step 6 : choose classifier

Set parameter

Final step : start to run

Finish

Thank you