文本挖掘技術(shù)在短信息過(guò)濾中的應(yīng)用研究.pdf

1、南昌大學(xué)碩士學(xué)位論文文本挖掘技術(shù)在短信息過(guò)濾中的應(yīng)用研究姓名：李楠申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：段隆振20081221ABSTRACTABSTRACTWithrapiddevelopmentofmobilecommunicationtechnologymobilephonetextmessagesbecomeallimportantwayofcommunicationofpeopleofallagesAtthesame

2、time，spammessagehasbecomeaproblemtroublingmobilephoneusersandtelecomoperatorsAtpresent，thespammessagefilteringtechnologyingeneralincludeswhiteandblacklisttechnologyrulesbasedfiltering，aswellasmatchingkeywordsContentbased

3、filteringisthemaintechnologytesolvespammessagesclasasificationInordertosolvethesetrouble，thisarticletextappliesminingtechnologytothemodelofmobilephonetofilterthesespammessagesInthispaperitdesignsandrealizesspammessagedue

4、l—filteringsystembasedonroughandKNNalgorithmThemodelincludesthepretreatment，featureextraction，weight；attributereduction，shortmessagefilteringandclassificationThisarticlecontainsthefollowingmainelements：1)Analyzingseveral

5、featureselectionalgorithmsandcomparingtheiradvantagesanddisadvantagesthroughexperiments(2)Proposingawayofcomputingtermswightbasedoninformationgainandvariance，elaboratingthismethod(3)Proposingadual—filteringmethodThecombi

6、nationofKNNandroughconstituteafilterRoughreductionalgorithmlowersvectorspacedemension，reducingthenumberoffeaturesandreducevectorspaceoftestmessageandenhancingthespeedofclassification(4)Thepretreamentofshortmessageinclude

7、soddityword；massnumbers；theintegrationofmessagecontentandpunctuation(5)Elebratingclassificationcriteria，assessingaclassificationtoolthroughexperiments，summingupresultofthestudyandpointingoutshortageandproposedsomeadvicet