資源描述:
《數(shù)據(jù)挖掘在電信企業(yè)客戶細(xì)分中的應(yīng)用》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、重慶大學(xué)碩士學(xué)位論文數(shù)據(jù)挖掘在電信企業(yè)客戶細(xì)分中的應(yīng)用姓名:程照星申請(qǐng)學(xué)位級(jí)別:碩士專業(yè):計(jì)算機(jī)技術(shù)指導(dǎo)教師:曾一;周曉蘭20041020煎慶大學(xué)硪士學(xué)位論文中文摘愛摘要夔著憊傣謄場(chǎng)競(jìng)爭(zhēng)垂奄不凝黧潮,驁簿鋅對(duì)不蕊靜客戶群實(shí)藏差異純落鏈纛暇務(wù),如何對(duì)客戶進(jìn)行細(xì)分和分類已成為當(dāng)前電信企業(yè)的迫切需求。本文討論了當(dāng)囂毫售金妲客戶縭分工終戇覡狀秘發(fā)震戇勢(shì),分孝廳了褥數(shù)據(jù)挖藏技術(shù)應(yīng)耀裂電信眾業(yè)中的必要性萃Ⅱ可行性,提出將數(shù)據(jù)挖掘聚類分析應(yīng)用到電倍企業(yè)的客戶細(xì)分王摻。數(shù)據(jù)挖掘(DataMining)就是從大擻的、不憲全的、有噪聲的、模糊的、隨耄陡熬實(shí)躲應(yīng)用數(shù)據(jù)中,提取戇含焱
2、其中夔、入鍛攀斃不熟遘戇、健又是潛在蠢建的信息和知識(shí)的:i過程。本文論述了數(shù)攝倉(cāng)庫(kù)秘?cái)?shù)援撼握的繁本理論;分板魄較了當(dāng)藏?cái)?shù)攥挖握領(lǐng)域通常采用的主要方法;深入研究了數(shù)據(jù)挖掘聚類分析中的K-Mean快速聚炭算法;繪燃將其逯饜到電信客戶細(xì)分的逑模過程。K.Mean快速漿類算法是在數(shù)據(jù)挖掘漿類算法中最常用的一種,它通過汁算樣本與初始中心點(diǎn)的距離,不斷調(diào)熬中心點(diǎn)的位置,使中心點(diǎn)到幫分樣本問的距離和達(dá)到最短,從而實(shí)現(xiàn)樣本的“聚集”和“分類”。結(jié)臺(tái)霪慶移動(dòng)通信有限責(zé)任公司的管理體制和業(yè)務(wù)模式,遵循CRISP—DM椽凇定義的商業(yè)數(shù)據(jù)挖掘過程,本文詳細(xì)論述了建模所涉及的商業(yè)理解、
3、數(shù)據(jù)理解、數(shù)據(jù)準(zhǔn)備、建立模型、模型評(píng)估和模型發(fā)布等過程。首先將客戶的各種消費(fèi)行為籀述數(shù)據(jù)德行K.Mean快速聚類,褥結(jié)合簾場(chǎng)業(yè)務(wù)的經(jīng)驗(yàn),將客戶按行為分為中余種組問差辮巨大,而組內(nèi)差異小的行為分組。最后在建立瀚電信企艟客戶繃分模鶩基礎(chǔ)±采用B/S模式魏計(jì)算祝體系結(jié)輸,結(jié)合軟件工程的思想,設(shè)計(jì)開發(fā)了“重慶移動(dòng)智能營(yíng)銷系統(tǒng)”,實(shí)現(xiàn)了數(shù)掇挖掘在積信金塹客戶纓分工作弱實(shí)際應(yīng)潮,實(shí)現(xiàn)了電信涎營(yíng)商針對(duì)不麗豹客戶群進(jìn)行差異化營(yíng)銷,差異化服務(wù)的目的,達(dá)到了提筒電信鹼業(yè)核心競(jìng)爭(zhēng)力的效果。關(guān)鍵詞:電信,數(shù)據(jù)挖掘,客戶細(xì)分,數(shù)據(jù)倉(cāng)庫(kù),聚類重慶大學(xué)碩士學(xué)位論文英文摘要ABSTRACTW
4、ithconstantaggravationofthetelecommunicationmarketcompetition,howtoimplementdifferencemarketingandservicetodifferentcustomersgroup,howtOsubdivideandclassi匆theactivedemandthathasalreadybecomepresenttelecommunicationenterprisestothecustomer.Thisarticlehasdiscussedthecurrentsituationso
5、fenterprise’Scustomeroftelecommunicationssubdividesanddevelopmenttrendsoftheworkatpresent,andhaveanalysedthatexcavatestechnologyofthedatatoapplytothenecessityandfeasibilityintelecommunicationenterprises,isitexcavatedataclusteranalysisapplytocustomer,telecommunicationofenterprisesubd
6、ivideandworktopropose.DataMiningisthetechniquethatrefinesthehidden,uJllcnownbutpotentialusefulinformationandknowledgefromhugemountof,incomplete,noisy,fuzzyandrandombusinessdata.11_liStexthasdescribedthebasic吐1eoriesindataWareb0LlSeanddatamining;Itisanalysedthatthemainmethodusedindat
7、aminingusuallyadoptsatpresent;TheK—Meanfastclustegsalgorithmincluster'sanalysisisdeeplystudied;Provideandapplyittomodelingcoursethattelecommunicationcustomersubdivides.K—Meanfastcluster'sNgofithmisthemostfrequentlyusedoneincluster’Salgorithminthedatamining,itcalculatesthedistancebet
8、weensamplesandiniti