Show simple item record

dc.contributor.authorRichardson, William Dale
dc.date.accessioned2014-07-01T04:30:26Z
dc.date.available2014-07-01T04:30:26Z
dc.date.issued2013-12
dc.identifier.otherrichardson_william_d_201312_ms
dc.identifier.urihttp://purl.galileo.usg.edu/uga_etd/richardson_william_d_201312_ms
dc.identifier.urihttp://hdl.handle.net/10724/30008
dc.description.abstractIn the field of machine learning, properties of data sets such as class imbalance and overlap often pose difficulties for classifier algorithms. A number of methods alleviate these difficulties by adjusting the distribution of the training data prior to classifier construction. Resampling is typically effected by weighting, removing, or duplicating instances, but finding a good resampling for the data set is a nontrivial problem. Genetic algorithms are frequently used to search for solutions in large, difficult search spaces. In this thesis, four evolutionary approaches are applied to the problem of instance resampling across a variety of data sets and classifier paradigms. In many cases, evolutionary pre-processing is able to produce better classifiers. In particular, an integer-based, one-to-one representation and a cluster-based, real-valued weighting encoding are shown to improve classifier performance on difficult data sets.
dc.languageeng
dc.publisheruga
dc.rightspublic
dc.subjectmachine learning
dc.subjectimbalance
dc.subjectundersampling
dc.subjectoversampling
dc.subjectinstance selection
dc.titleEvolutionary instance resampling for difficult data sets
dc.typeThesis
dc.description.degreeMS
dc.description.departmentArtificial Intelligence
dc.description.majorArtificial Intelligence
dc.description.advisorKhaled Rasheed
dc.description.committeeKhaled Rasheed
dc.description.committeeWalter D. Potter
dc.description.committeePrashant Doshi


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record