Abstract: We propose a methodology for predictive classification from gene expression data, able to combine the robustness of highdimensional statistical classification methods with the comprehensibility and interpretability of simple logic-based models. W
RELATIONALSUBGROUPDISCOVERYFORGENE
EXPRESSIONDATAMINING
ˇˇep´F.Zelezn´y ,J.Tolar ,N.Lavraˇc andO.Stˇankov´a
GerstnerLaboratory,DepartmentofCybernetics,
CzechTechnicalUniversity,Technick´a2,
16627Prague,CzechRepublic
DivisionofHematology,Oncology,BloodandMarrowTransplant,Departmentof
Pediatrics,UniversityofMinnesotaMedicalSchool420DelawareStreet,55455Minneapolis,USA
DepartmentofKnowledgeTechnologies,JoˇzefStefanInstitute
Jamova39,1000Ljubljana,Slovenia
and
NovaGoricaPolytechnic,Vipavska13,5000NovaGorica,Slovenia
zelezny@fel.cvut.cz,tolar003@umn.edu,nada.lavrac@ijs.si,step@labe.felk.cvut.cz
Abstract:Weproposeamethodologyforpredictiveclassi cationfromgeneexpressiondata,abletocombinetherobustnessofhigh-dimensionalstatisticalclassi cationmethodswiththecomprehensibilityandinterpretabil-ityofsimplelogic-basedmodels.We rstcon-structarobustclassi ercombiningcontribu-tionsofalargenumberofgeneexpressionval-ues,andthensearchforcompactsummariza-tionsofsubgroupsamonggenesassociatedintheclassi erwithagivenclass.Thesubgroupsaredescribedbymeansofrelationallogicfea-turesextractedfrompubliclyavailablegeneannotations.Thecurseofdimensionalityper-tainingtothegeneexpressionbasedclassi- cationproblemduetothelargenumberofattributes(genes)isturnedintoanadvantageinthesecondarysubgroupdiscoverytask,asheretheoriginalattributesbecomelearningexamples.Introduction
Manytasksofautomatedknowledgediscoveryfromgeneexpressionmicroarraydatabydatamin-ingalgorithmsaimatconstructingclassi ersabletodiagnoseacancertypefromageneexpressionpro le.Seeeg.theseminalpapers[1,2]forreference.
Thistaskischaracterizedbytheabundanceofattributes(eg.simultaneouslymeasuredgeneexpres-sionvalues)ononehandconfrontedwiththeshort-ageoftheavailablesamples(eg.patients/tissuessub-jecttomeasurements)ontheotherhand.Itisknownfromscienti cdiscoveryliteraturethatsuchdomainsarepronetoover tting:over ttedclassi ersarechar-acterizedbysigni cantlydecreasedpredictiveaccu-racyonunseensamplescomparedtothetrainingsetaccuracy.
Informally,indomainscharacterizedbyasmallnumberofexamplesandalargenumberofat-tributes,over ttingoccursbecausesomeartefacts(“ ukes”)ofactuallyirrelevantattributecombina-tionscanemergesimplybymeansofchanceandappearsigni cantwithrespecttotheexamplesavail-abletothemachinelearningalgorithm.
Toavoidtheover ttingpitfall,state-of-artap-proachesconstructcomplexclassi ersthatcombinerelativelyweakcontributionsofuptothousandsofgenes(attributes)toclassifyadisease.Real-valuedsupportvectormachine(SVM)[3]modelsarecur-rentlyspeci callypopularinthegeneexpressiondataminingdomain.However,classi ersbasedonmanyreal-valuedattributeshaveanimportantdraw-back:theyarenotappropriateforexpertinterpreta-tion.Thecomplexityofsuchclassi erslimitstheirtransparencyandconsequentlythebiologicalinsighttheycanprovide.Althoughitispossibletoextracttheattributeswithmaximalcontributionweight,thelogicalconnectionsamongtheextractedattributesarethenlost.
Inourpreviouswork[4],wehavetestedthefeasibilityofconstructingsimpleyetrobustlogic-basedclassi ers,intheformofpropositionalrulesamenabletodirectexpertinterpretation,byanin-novativealgorithmicmethodologyofsubgroupdis-covery.Suchrulestypicallyincludetwoto vegeneexpressionattributesand,incontrasttomarkersobtainedfromSVMschemes,theserulesexplicitlystresstheimportanceofthecorrelationoftheactiv-ity(ornon-activity)ofanarrowgeneset.Followingourwork,furtherpapersappeared[5,6]proposingmethodsforconstructingverysimple,interpretable
Data mining for gene exp... 16页 免费 Relational subgroup disc... 暂无评价 6页 免费 Gene Expression Data Min... 8页 免费 An evolutionary clusteri...
Relational data mining w... 暂无评价 19页 免费 Subgroup discovery with ... Adaptive Gene Expression... 5页 免费 Mining Predicate Associa... 暂无评价...
Mining Heterogeneous Gen... 暂无评价 7页 免费 Relational subgroup disc..... Gene Expression Data Mining for Functional Genomics using Fuzzy Technology ...
Relational Descriptive Analysis of Gene Expression Data STAIRS-2006 (3rd ... Saarbruecken 2004 Advanced course on Knowledge Discovery and Data Mining, ...
Discovery of first-order... 暂无评价 8页 免费 Relational Data Mining a... 暂无评价 12页 免费 Relational Pattern and S... 暂无评价 12页 免费 ...
Relational subgroup disc... 暂无评价 6页 免费 Gene expression Query-dr.... Combining location and expression data for principled discovery of genetic ...
Relational Data Mining a... 暂无评价 12页 免费 A methodology for knowle... Expert-Guided Subgroup Discovery Methodology and Application This paper ...
Relational subgroup disc... 暂无评价 6页 免费... Comparison, Class Prediction or Class Discovery. ... analysis of gene expression microarray Data. ...
Given the style of data needed for link discovery, pattern learning for link discovery requires relational data mining. The most widely studied methods for...
Relational subgroup disc... 暂无评价 6页 免费 Analysis of Microarray G.... Gene expression Query-driven module discovery in microarray data Motivation: ...
热门文档