發(fā)布時(shí)間：2018-03-04 00:19

  本文選題：數(shù)據(jù)挖掘　切入點(diǎn)：高端用戶　出處：《浙江工商大學(xué)》2013年碩士論文　論文類型：學(xué)位論文

【摘要】：隨著通信行業(yè)的發(fā)展、手機(jī)的普及,我國電信行業(yè)的競(jìng)爭(zhēng)越來越激烈。電信行業(yè)的經(jīng)營理念由原來的以產(chǎn)品為中心轉(zhuǎn)變?yōu)橐钥蛻魹橹行�。電信行業(yè)通過擴(kuò)大手機(jī)用戶來提高自己的業(yè)務(wù)已經(jīng)非常困難。目前,對(duì)運(yùn)營商來說最重要的是保留用戶,尤其是高價(jià)值用戶,避免用戶流失。保留用戶的首要前提是了解用戶。 首先,本項(xiàng)目研究如何識(shí)別電信公司高端用戶。根據(jù)電信公司業(yè)務(wù)人員對(duì)高端用戶的商業(yè)理解大概確定高端用戶應(yīng)該滿足的條件,在此基礎(chǔ)上使用數(shù)據(jù)挖掘方法具體確定哪些用戶為高端用戶,在此基礎(chǔ)上分析高端用戶應(yīng)該滿足的特征。在分析高端用戶時(shí),本項(xiàng)目從兩個(gè)角度進(jìn)行了研究：業(yè)務(wù)角度和統(tǒng)計(jì)角度。最后合并兩種方法得到的結(jié)果確定最終的高端用戶。 其次,本項(xiàng)目研究流失用戶識(shí)別。隨著電信行業(yè)競(jìng)爭(zhēng)加劇,沒有一個(gè)運(yùn)營商可以肯定其用戶不會(huì)轉(zhuǎn)向其他運(yùn)營商。因此,預(yù)測(cè)哪些用戶將會(huì)流失是運(yùn)營商保留用戶的前提。本項(xiàng)目在理解業(yè)務(wù)的基礎(chǔ)上,先對(duì)流失用戶進(jìn)行界定,然后用決策樹算法進(jìn)行建模,并對(duì)算法CR、 QUEST、C5.0和基于Boosting的C5.0得到的結(jié)果進(jìn)行比較,從中選擇基于Booting的C5.0作為本項(xiàng)目最終使用的模型。 最后,從被預(yù)測(cè)為將會(huì)流失的用戶中選出高端用戶,作為運(yùn)營商挽留的重要用戶。在此基礎(chǔ)上,分析每類被預(yù)測(cè)為流失的高端用戶挽留措施。
[Abstract]:With the development of the communication industry, the popularity of mobile phones, The competition of telecommunication industry in our country is more and more intense. The management idea of telecommunication industry has changed from product center to customer center. It is very difficult for telecom industry to improve its business by expanding the number of mobile phone users. At present, The most important thing for operators is to keep users, especially high-value users, to avoid the loss of users. First of all, this project studies how to identify high-end users of telecom companies. According to the business understanding of high-end users of telecom companies, we can determine the conditions that high-end users should meet. On this basis, we use data mining method to determine which users are high-end users, and then analyze the characteristics that high-end users should satisfy. This project has carried on the research from two angles: the business angle and the statistical angle. Finally, the result of combining the two methods to determine the final high-end users. Secondly, this project studies the loss of user identification. As competition in the telecommunications industry intensifies, no operator can be sure that its users will not turn to other operators. Predicting which users will be lost is a prerequisite for operators to retain users. Based on the understanding of the business, the project defines the lost users first, and then uses the decision tree algorithm to model the users. The results obtained from the algorithm CR-QUESTC5.0 and C5.0 based on Boosting are compared, and C5.0 based on Booting is selected as the final model of this project. Finally, the high-end users are selected from the users who are predicted to be lost, as the important users to be retained by the operators. On this basis, the retention measures of each type of users are analyzed.
【學(xué)位授予單位】：浙江工商大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2013
【分類號(hào)】：TP311.13;F626

【參考文獻(xiàn)】

相關(guān)期刊論文 前4條

1 朱浩剛,孫煜鷗,戴偉輝;基于數(shù)據(jù)挖掘的移動(dòng)通訊業(yè)客戶流失管理[J];計(jì)算機(jī)工程與應(yīng)用;2004年01期

2 趙興華;李杰;王云峰;;數(shù)據(jù)挖掘在電信中的應(yīng)用分析[J];科技情報(bào)開發(fā)與經(jīng)濟(jì);2007年12期

3 吳志勇;戴曰章;鞠傳香;;數(shù)據(jù)挖掘在電信客戶流失中的應(yīng)用[J];山東理工大學(xué)學(xué)報(bào)(自然科學(xué)版);2007年05期

4 李晶;;淺談針對(duì)消費(fèi)者行為的企業(yè)營銷策略[J];現(xiàn)代經(jīng)濟(jì)信息;2010年04期

，

本文編號(hào)：1563326