本文選題：數(shù)據(jù)挖掘 + 客戶細分模型��； 參考：《北京郵電大學(xué)》2010年碩士論文

【摘要】：企業(yè)建立自己的數(shù)據(jù)庫系統(tǒng),由計算機管理代替手工操作,以此來收集、存貯、管理業(yè)務(wù)操作數(shù)據(jù),改善辦公環(huán)境,提高操作人員的工：作效率。企業(yè)需要從業(yè)務(wù)數(shù)據(jù)中提取有用的信息,幫助他們在業(yè)務(wù)管理和發(fā)展上作出即時、正確的判斷。 當(dāng)一個企業(yè)明確了自己的客戶之后,緊接著就應(yīng)該做客戶細分。不同的客戶對同一產(chǎn)品的需要存在著明顯的差別,客戶對產(chǎn)品的要求越來越理性和嚴格,對企業(yè)服務(wù)的整體質(zhì)量也提出了更高要求。不同類型的客戶選擇的往往不僅是產(chǎn)品的單一特性,還可能是產(chǎn)品特性的某種組合。對于企業(yè)來說,同一產(chǎn)品不可能滿足市場上所有的客戶的需要,只能面向某一種類型的客戶。另一方面,某一一特定的產(chǎn)品要不僅滿足單一類型的客戶,還要滿足多范圍、多層次、有著不同需要的客戶群。 客戶細分的目的,就是要更精確地回答誰是我們的客戶,客戶到底有哪些實際需要,企業(yè)應(yīng)該去吸引哪些客戶,應(yīng)該重點保持哪些客戶,應(yīng)該如何迎合重點客戶的需求等重要問題。 企業(yè)可以通過響應(yīng)率分析能夠有效的降低市場推廣的費用,同時能夠更加有針對性的面對目標(biāo)市場,達到以最小的投入獲得最佳效果的目的。需要構(gòu)建預(yù)測模型,找到最合適的響應(yīng)客戶,預(yù)測哪些客戶能夠響應(yīng),以及響應(yīng)的可能性是多少。 針對廣告營銷市場的不斷發(fā)展,企業(yè)收集了大量的客戶資料。數(shù)據(jù)挖掘需求來自新雅迪傳媒。為了便于廣告中心制定較為合理的營銷策略,將用SPSS Clementine建立模型,以提升新客戶開發(fā)的成功率,降低長單客戶的流失率。 本文根據(jù)業(yè)務(wù)部門需求,經(jīng)過與業(yè)務(wù)人員的不斷溝通,將營銷過程、客戶信息與數(shù)據(jù)挖掘技術(shù)相結(jié)合,經(jīng)過數(shù)據(jù)理解、數(shù)據(jù)清洗、模型訓(xùn)練等數(shù)據(jù)挖掘過程,設(shè)計并實現(xiàn)了客戶細分模型和客戶響應(yīng)預(yù)測模型。再對模型進行評估和部署,有效的從雜亂無章的客戶數(shù)據(jù)中發(fā)現(xiàn)具有商業(yè)價值的信息。
[Abstract]:In order to collect, store and manage the operation data, improve the office environment and improve the working efficiency, the enterprise establishes its own database system and uses computer management instead of manual operation to collect, store and manage the operation data. Enterprises need to extract useful information from business data to help them make immediate and correct judgments on business management and development. When an enterprise has identified its own customers, the next step should be customer segmentation. Different customers have different needs for the same product. The requirements of customers for the same product are more and more rational and strict, and the overall quality of the enterprise service is also put forward higher requirements. Different types of customers often choose not only the single feature of the product, but also some combination of the product characteristics. For enterprises, the same product can not meet the needs of all customers in the market, only for one type of customers. On the other hand, a particular product should not only meet a single type of customer, but also meet a multi-scope, multi-level, with different needs of the customer base. The purpose of customer segmentation is to answer more precisely who our customers are, what the actual needs of our customers are, what customers enterprises should attract, and which customers should be kept. How to meet the needs of key customers and other important issues. Through the response rate analysis, enterprises can effectively reduce the cost of market promotion, at the same time, they can face the target market more pertinently, and achieve the goal of getting the best effect with the minimum investment. A predictive model is needed to find the most appropriate response customer, which customers can respond, and what is the likelihood of the response. In view of the continuous development of advertising marketing, enterprises have collected a large number of customer information. The demand for data mining comes from New Yadi Media. In order to facilitate the advertising center to formulate more reasonable marketing strategy, SPSS Clementine will be used to establish a model to improve the success rate of new customer development and reduce the loss rate of long single customer. According to the demand of business department, this paper combines marketing process, customer information with data mining technology, data understanding, data cleaning, model training and other data mining processes through continuous communication with business personnel. The customer subdivision model and customer response prediction model are designed and implemented. Then the model is evaluated and deployed, and the information of commercial value is found from the chaotic customer data.
【學(xué)位授予單位】：北京郵電大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2010
【分類號】：TP311.13

【參考文獻】

相關(guān)期刊論文 前2條

1 馮天宇;何衛(wèi)平;方偉;杜承烈;;基于數(shù)據(jù)挖掘的CRM系統(tǒng)的研究[J];科學(xué)技術(shù)與工程;2007年15期

2 張?zhí)咸?丁莉,王景光;數(shù)據(jù)挖掘——CRM中的動力[J];企業(yè)技術(shù)開發(fā);2003年06期

相關(guān)碩士學(xué)位論文 前1條

1 董云峰;基于數(shù)據(jù)挖掘的CRM系統(tǒng)研究與設(shè)計[D];山東大學(xué);2007年

，

本文編號：1926051