PDF《朝阳科技大学》

朝阳科技大学 资讯管理系 硕士论文 云端市集中顾客需求分类之研究 The Study of Customer Requirement Categorization in a Cloud Market 指导教授:王淑卿 博士 研究生:林雨农 中华民国

102 年3月25 日 朝阳科技大学资讯管理系 Department of Information Management Chaoyang University of Technology 硕士论文 Thesis for the Degree of Master 云端市集中顾客需求分类之研究 The Study of Customer Requirement Categorization in a Cloud Market 指导教授:王淑卿 博士(Shu-Ching Wang) 研究生:林雨农(Yu-Nung Lin) 中华民国

102 年3月25 日March 25,

2013 I 摘要 在资讯日新月异的时代下,电子商务的蓬勃发展已成为一股趋势,也 改变了人们的消费习惯,使企业越来越重视网路商店的区块和商机,并愿 意在此投入成本提供更好的服务;

而网际网路的快速发展,频宽的进步, 也使得云端运算(Cloud Computing)这个概念迅速的窜起.

由於云端运算的 特性,使得现今企业在网路上提供的服务越来越多元化,也使云端环境的 使用者快速的增加,改变了现有的商业模式,预期未来电子商务将与云端 运算有密不可分的关系. 由於云端运算的环境非常广大,因此大量的云端服务供应商和使用者 在此进行交易,使得云端使用者在云端市集上,面临了难以迅速找到提供 适当服务的云端服务供应商.所以,如何能保障云端运算环境中的使用者 能迅速找到适当的服务需求是云端服务必须探讨的议题.因此,本研究提 出云端市集的商业模式层级及订阅机制来减少云端使用者所要寻找的资讯 量,让云端服务的使用者能够享有更好的服务品质. 再者,因为云端市集中拥有大量的云端使用者,使得云端服务供应商 在面对大量云端使用者所提出的服务需求,难以迅速找到提供适当需求的 服务.因此,依贫耸褂谜吖サ姆裥枨蠹吐冀掷唷⑵饰觥 且以自动化的方式找出云端使用者正确的服务需求,是本研究的主要重 II 点.本研究运用决策法则的推导(Decision Rule-based Induction) 及LLSF(Linear Least Square Fit)二阶段的方法,撷取云端使用者所提出的服 务需求之特性.研究中利用需求关键字的特性和需求出现的频率为基础来 判断需求的类别,以撷取正确的服务需求.透过本研究所提出的方法,可 以快速的回应云端使用者所提出的服务需求,不仅可以减少人力成本,更 可提高云端服务的品质与效能,准确且快速的自动给予云端使用者适当的 服务,提高顾客的满意度. 关键词:云端运算、云端服务、电子市集、商业模式、订阅机制、文件分 类、决策法则推导、LLSF III Abstract Because of the rapidly changing in information, the E-Commerce has become a trend, but also changed people'

s consumption habits. To enable enterprises to more and more attention to the block and opportunities of the online store, and are willing to provide better input costs services;

the rapid development of Internet bandwidth of progress, but also makes the concept of Cloud Computing spring up rapidly. Due to the characteristics of cloud computing, making the services provided by the companies on the Internet today are more and more diversified, so the users in the cloud environment are rapid increase, and the users will change the existing business model, expected future e-commerce and cloud computing are inseparable relationship. The cloud computing environment is very large, so a lot of cloud service providers and users in here to make transactions, so that will make the cloud users to face with a difficult problem to quickly find the cloud service providers to provide appropriate services in the cloud marketplace. Therefore, how to protect the cloud users in the cloud-computing environment, make them can quickly find the appropriate service of needs is the necessary issues of cloud services. Therefore, this study present the level of the business model of the cloud marketplace and subscription mechanisms to reduce the information amount of cloud users are looking for, and so them can able to enjoy a better quality of service. Furthermore, because there are a lot of cloud users in the cloud marketplace, so the cloud service providers must have the problems of facing the large service needs from a large numbers of cloud users, it is difficult to quickly find the appropriate needs of the service. Therefore, based on the past records of the service needs of the cloud users will be sorted, analyzed and automated way to identify the demand for services of cloud users, that is the objects of this study. In this study, we use the decision rules-based induction (Decision Rule-based Induction) and LLSF (Linear Least Square Fit) two-stage method to retrieve the characteristics of the demand for services of cloud user. Demand keyword characteristics and needs of the frequency-based study to determine the type of demand in order to retrieve the correct service needs. Through the method proposed in this study, a rapid response to the cloud user demand for services, not only can reduce labor costs, but also enhance the quality and performance of cloud services. It will automatic IV given cloud users to appropriate services accurate and fast, so that can improve customer satisfaction. Keywords: Cloud Computing, Cloud Service, e-Market, Business Model, Publish Subscription, Document classification, the decision rule base, LLSF V 谢志 这本硕士论文的完成,真的是很不可思议的事情!由於我本身大学时 期是念中文系,而且也工作了一段时日,最后选择要以日间部的学生身分 回学校来完成资管硕士学位,这一路走来真的是很辛苦.当然,要完成这 个艰的工作,完全都是来自於我身边周遭许多鼓励我以及支持我的老 师、朋友以及家人,在此,请让我对各位献上十二万分的谢意. 首先,最要感谢的就是我的指导教授王淑卿老师!淑卿老师不嫌弃我 非本科系出身,愿意收我为学生,并且总是不厌其烦的指导我,常让我感 激万分.而我也因为毫无资讯背景基础,导致在学业上常常力不从心,再 加上有家庭跟经济上的压力,几乎快支撑不住,甚至萌生休学的想法;