【文章內(nèi)容簡(jiǎn)介】 It involves three steps. In the ?rst step, we use fuzzy clustering to categorize customers. A key feature of this fuzzy clustering model is that the number of groups is determined automatically from data using the partition entropy as a validity measure. In the second step, the dimension (or number of attributes) for each cluster (or group of customers) is reduced by selecting only the most informative attributes. Selection is based on the information loss of an attribute。 a quantity puted using the entropy of the attribute, and that of the whole group (or cluster). Consequently, and as a result to this second step, we obtain a set of groups of customers each of them described by a distinct set of attributes judged as being the most informative. In the third and ?nal step of our approach, each reduced cluster is trained by a feedforward backpropagation work to extract useful knowledge. Hence, we obtain a set of backpropagation works each encoding in its connections a customer pro?le (or model) and which could be used subsequently in classifying new and previously unknown customers. The rest of this paper is structured as follows. In Sec. 2, we detail our model. Section 3 presents experimental analysis。 while the last section o?ers concluding remarks and shades light on future research. Our Proposal Our approach is threestep and is summarized in Fig. 1. First, we use fuzzy 8 clustering to identify groups of customers. An important characteristic of our approach is that the number of clusters is puted automatically using the partition entropy as a validity measure. In this way, the optimal number of clusters is that producing the lowest partition entropy. As a result to this ?rst step, we obtain a set of groups each describing a customer category whose consuming habits, ines, etc., are similar. Second, we proceed to a dimensionality reduction retaining only pertinent attributes for each cluster. In fact, within a group of customers (cluster), not all original attributes are pertinent (or informative) and some of them should be discarded. Here again, we developed an entropybased approach to decide whether or not an attribute is important with regard to a given group of customers using a userde?ned threshold for the information loss. Third, each cluster is fed into a backpropagation neural work in order to extract useful knowledge. In this way, and as a result to the whole approach, we obtain a set of customer pro?les each encoded within the connections of a backpropagation work. Hence, classi?cation (categorization) of new, previously unknown customers bees an easy task. In the following subsection, we will detail each step of our approach. . Second step: Attribute selection Naturally, within the same cluster (the same group of customers), not all original attributes are “informative” and some of them should be discarded. As a measure for the “information” conveyed by an attribute, we will use the entropy based on the frequency of the values taken by that attribute within a given cluster. In order to pute this frequency, we will use the same clustering approach in Fig. 2 but applied to a data set posed of the values taken by that attribute. This is motivated by the following. The number of possible values for an attribute within a given a group of customers could not be puted exactly especially for numerical attributes where values could be “very close”, but not exactly the same. By applying the clustering approach in Fig. 2 to a given attribute within a given group of customers, we obtain a set of clusters each embedding values for that attribute that are very similar。 and therefore considered as being a single value occurring as many times as the size of the 9 corresponding cluster. Before going into further details, now we need the following notation. 3 外文文獻(xiàn)譯文（ 2） 客戶的知識(shí)關(guān)系管理：整合知識(shí)管理和客戶關(guān)系管理過(guò)程 摘要 市場(chǎng)的激烈的競(jìng)爭(zhēng)和業(yè)務(wù)環(huán)境的快速變化，信息利用已經(jīng)成為企業(yè)增強(qiáng)競(jìng)爭(zhēng)優(yōu)勢(shì)的關(guān)鍵，知識(shí)管理 (KM)和客戶關(guān)系管理 (CRM)過(guò)程是一個(gè)嶄新的研究領(lǐng)域，但是，圍繞它科學(xué)研究和相關(guān)文獻(xiàn)仍然有限，另外，客戶獲取、保持、擴(kuò)張過(guò)程中知識(shí)管理的作用提高客戶滿意仍然處于研究和報(bào)告水平。本論文的目的根據(jù)知識(shí)管理和客戶關(guān)系管理不同的模型，結(jié)合知識(shí)管理和客戶 關(guān)系管理提供一個(gè)概念性的框架，這個(gè)框架稱為客戶知識(shí)關(guān)系管理。主要強(qiáng)調(diào)了概念為基礎(chǔ)的客戶知識(shí)的概念 (了解客戶，客戶信息，客戶知識(shí) )。因此，本文研究知識(shí)管理過(guò)程的發(fā)展（客戶知識(shí)發(fā)現(xiàn)、處理、運(yùn)用）。本文分析研究喬丹公司如何利用知識(shí)管理過(guò)程提高客戶關(guān)系管理過(guò)程。根據(jù)該公司的數(shù)據(jù)采集，結(jié)果分析表明知識(shí)管理過(guò)程對(duì)客戶關(guān)系管理過(guò)程產(chǎn)生了積極的影響。 關(guān)鍵詞：知識(shí)獲取、知識(shí)創(chuàng)造、知識(shí)應(yīng)用；客戶的獲取、客戶保留、客戶挖掘。 介紹 由于客戶知識(shí)革命的快速發(fā)展，建立同客戶高效且有效的的關(guān)系非常需要知識(shí)管理過(guò)程。另外，客戶關(guān)系管理的本質(zhì)是知識(shí)管理，因?yàn)榭蛻絷P(guān)系管理幫助企業(yè)加強(qiáng)服務(wù)，快速響應(yīng)客戶需求。企業(yè)需要加強(qiáng)與客戶的互動(dòng)，確定知識(shí)管理的相關(guān)活動(dòng)領(lǐng)域，改善流程。 此外，知識(shí)管理是一個(gè)捕獲、創(chuàng)建和應(yīng)用知識(shí)使客戶關(guān)系管理過(guò)程成功的方法。更進(jìn)一步， Gebert 提出客戶關(guān)系管理和知識(shí)管理已經(jīng)在商業(yè)市場(chǎng)中引起廣泛的興趣。這兩種方法關(guān)注分配資源，支持業(yè)務(wù)活動(dòng)，以獲得競(jìng)爭(zhēng)優(yōu)勢(shì)，盡管兩個(gè)概念實(shí)際上作為分開(kāi)的研究的領(lǐng)域。 Lin 認(rèn)為客戶關(guān)系管理和知識(shí)管 理對(duì)每個(gè) 10 市場(chǎng)決策者和信息技術(shù)專業(yè)人員有重要意義。 客戶關(guān)系管理的知識(shí)管理是獲得客戶滿意重要方法。知識(shí)管理過(guò)程和客戶關(guān)系管理過(guò)程的結(jié)合一個(gè)新的研究領(lǐng)域，因此圍繞它科學(xué)研究和相關(guān)文獻(xiàn)仍然有限，知識(shí)管理過(guò)程對(duì)客戶關(guān)系管理過(guò)程中的影響利用仍然處于研究和報(bào)告水平。 本論文的目的在于提出客戶知識(shí)關(guān)系管理概念模型，整合知識(shí)管理過(guò)程和客戶關(guān)系管理過(guò)程提提提高客戶滿意度。將會(huì)實(shí)現(xiàn)下面的目的： 確定目前企業(yè)如何處理客戶，通過(guò)分析企業(yè)的使命完成。 提高對(duì)客戶知識(shí)的獲取以獲得新客戶，通過(guò)利用客戶知識(shí)保持現(xiàn)有客戶，拓展客戶知識(shí) 擴(kuò)大與客戶的關(guān)系。 描述未來(lái)企業(yè)如何處理客戶關(guān)系，識(shí)別企業(yè)的愿景實(shí)現(xiàn)目標(biāo)。 接下來(lái)部分，查閱相關(guān)文獻(xiàn)，第三章提出客戶知識(shí)關(guān)系管理過(guò)程模型，第四章提出研究方法。 文獻(xiàn)綜述 本節(jié)概述不同文獻(xiàn)關(guān)于知識(shí)管理過(guò)程，提供了 CRM 過(guò)程的描述。最后，描述相關(guān)文獻(xiàn)對(duì)知識(shí)管理過(guò)程和客戶關(guān)系管理過(guò)程的關(guān)系。 知識(shí)管理過(guò)程 知識(shí)管理的目的不是處理所有知識(shí)，對(duì)企業(yè)來(lái)說(shuō)管理知識(shí)是最重要的。它涉及應(yīng) 用收集的知識(shí)和全部人力物力實(shí)現(xiàn)企業(yè)特定的目標(biāo)，合適的人利用合適的信息在合適的時(shí)間幫助人們收集和分析知識(shí)提 高企業(yè)效益。 作者提出并發(fā)展了一個(gè)概念和知識(shí)管理清晰模型?；谥R(shí)管理文獻(xiàn)中多種模型的調(diào)查，從知識(shí)的過(guò)程中捕獲知識(shí)，知識(shí)處理中產(chǎn)生所需的知識(shí)，知識(shí)提取中應(yīng)用知識(shí)。根據(jù)下面表 1 中知識(shí)管理過(guò)程的分類法。 客戶關(guān)系管理過(guò)程 客戶關(guān)系管理是近年來(lái)相關(guān)領(lǐng)域中出現(xiàn)的最熱門的課題，因?yàn)槠髽I(yè)客戶關(guān)系管理的價(jià)值。更進(jìn)一步，客戶關(guān)系管理成為所有企業(yè)的重要的業(yè)務(wù)流程和任務(wù)。 作者提出對(duì)客戶關(guān)系管理的清晰模型概念化，如下圖表 2依靠 CRM流程的分類法，客戶流程獲取新客戶，客戶流程保持客戶，流程擴(kuò)大與客戶的關(guān)系。 知識(shí)管理和客戶關(guān)系管理流程之間的關(guān)系 客戶關(guān)系管理中的知識(shí)管理很重要，因?yàn)檫@幫助企業(yè)做更好的服務(wù)，加強(qiáng)產(chǎn) 11 品的質(zhì)量，減少費(fèi)用和迅速響應(yīng)客戶。然而，企業(yè)管理知識(shí)最突出的挑戰(zhàn)是在所有不同部門成員之間計(jì)算和整合知識(shí)。因此，知識(shí)管理是成功的客戶關(guān)系管理戰(zhàn)略關(guān)鍵因素之一，提高服務(wù)質(zhì)量，降低服務(wù)成本，發(fā)展新產(chǎn)品和服務(wù)客戶。只有少數(shù)的企業(yè)實(shí)現(xiàn)把信息轉(zhuǎn)成客戶知識(shí)。 另外， Salomann 區(qū)分三種知識(shí)流動(dòng)在企業(yè)和客戶交互中的起到至關(guān)重要的作用：客戶知識(shí)在他們的購(gòu)買周期間支持客戶；一個(gè)連續(xù)的知識(shí)流直接從公 司向它的客戶。來(lái)自客戶的知識(shí)的必須合并公司產(chǎn)品和服務(wù)的創(chuàng)新和發(fā)展。關(guān)于客戶知識(shí)的收集通過(guò)客戶關(guān)系管理服務(wù)，支持流程和客戶關(guān)系管理分析過(guò)程。 Ocker 和 Mudambi 指出這些企業(yè)需要探索和改進(jìn)客戶關(guān)系管理和知識(shí)管理的方法為企業(yè)和客戶增加額外的知識(shí)價(jià)值。