基于Inclusion-Exclusion原理的頻繁項(xiàng)集壓縮方法的研究.pdf

1、Y5≤G9l後璺大學(xué)碩士學(xué)位論文學(xué)校代礬：10246學(xué)號(hào)：032021170基于InclusionExclusion原理的頻繁項(xiàng)集壓縮方法的研究院專姓系：計(jì)算機(jī)與信息技術(shù)系業(yè)：計(jì)算機(jī)軟件與理論名：楊曉明指導(dǎo)教師：張守志副教授完成日期：2006年5月6日●●●●●早十Inclusion—Exclusion蟓胖的頻繁項(xiàng)集件縮方注的研究AbstractFrequentitemsetsminingisanessentialstepinmanyd

2、ataminingtasksHoweversincethenumberofpotentialfrequentitemsetsisgrowsexponentiallywiththenumberofitems，thescaleoffrequentitemsetsisusuallyhuge，whichnotonlyaffectstheefficiencyofminingtask，butalsestricttheeffectofminingta

3、skbecausesuchalargesetmakeitdifficultfortheusertounderstandandprocess，thusrestricttheapplicationanddataminingInfact，thereismuchredundancyinthecollectionoffrequentitemsetswiththeirsupportinformationUsingthewell—knownInclu

4、sion—ExclusionformulaoftheSetTheorywecandiscovertheredundancyandalsocaninfertheredundancyinformationSotheredundancypartcanbecutfromthecollectionandthusreducethecollectionsizeTherearemanyparticularwaytodothisreductionever

5、ywaycanbecalledacondensedrepresentationByminingsomekindsofcondensedrepresentation，wecancutdownthecollectionsizeeffectivelyThesizeofcondensedrepresentationisusuallyfarsmallerthentheoriginalonewhilekeepsalmostallinformatio

6、nMainworkofthisthesisincludes：奪Weproposedtheconceptofstrongrule，andweproposeanewcondensedrepresentationbycombiningfrequentshosteststrongrulesandfrequentitemsets，whichgiveabetterandmoreconciseapproximationoftheoriginalcol

7、lection奪Weintroducedtheconceptofalmostderivableitemsetsandshowthecollectionofallnonalmostderivableitemsetscanberegardedasacondensedrepresentationofthecollectionoffrequentitemsets奪Wediscussedmostknowncondensedrepresentati

8、on，andshowtheirrelationshipwiththeInclusion—ExclusiontheoryWealsoextendedtheseveralknowncondensedreprensentationintwoways：approximationandstrongrulesThenweintroducedthekalmost—freenessframework，andgiveanin—depthanalysisa

9、bouttherelationsbetweentheframeworkandseveralcondensnedreprenesntationsThenwegavetwoeffectivetunablecondensedreprensentaionbasedonkalmost—freenessframeworkKeyword：Datamining，F(xiàn)requentItemset，CondensedRepresentation，Inclus