基于前景理论的B2C消费者行为决策模型研究

发布时间：2018-02-28 21:05

本文关键词： 决策模型前景理论 B2C电子商务网络口碑　出处：《华南理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着电子商务的蓬勃发展,网络消费者表现出了愈来愈多的与以往不同的行为特征。如:随着“双11”活动越来越被消费者期待和认同,在活动开展的过程中,存在大量网络消费者熬夜刷单、冲动消费(“剁手族”)、节后大量退货等行为。这些行为在一定程度上反映了网络消费者行为决策过程中的“有限理性”甚至是“非理性”的特点。因此,针对当前网络消费者的行为特征,本研究的目的在于:能够合理地对网络购物消费者“有限理性”或“非理性”行为决策过程进行正确的建模,从而为企业开展运营管理及服务营销管理提供切实可行的建议。本研究主要内容包括:(1)网络购物情境下的消费者行为决策研究及建模。运用服务蓝图法确定服务流程各阶段,将B2C电子商务服务流程分为决策、通达、交易、支付、物流和售后6个子流程阶段;针对各个流程中的顾客行为决策建立价值函数、权重函数及前景函数;通过引入直觉模糊数以完善数学模型;最后,通过定义多属性决策问题,确定了京东、1号店和当当为备选方案集及相对应的决策指标属性的结构。(2)基于网络口碑的决策模型定量化。在已建立的行为决策模型的基础上,通过挖掘网络口碑定量表示决策模型。首先,使用java编程语言,按照服务流程进行网络口碑的分类,实现了基于SVM算法的网络口碑服务流程分类模型;分类完成后,调用weka数据分析工具,对类内的网络口碑进行聚类分析。主要运用EM算法,确定了类内聚类的簇数(即指标数),并辅助QFD调研方法确定各个类中子决策指标权重。经过聚类后,各个子流程共得到13个决策指标;进一步使用ROST情感分析工具对网络口碑进行情感评分,依据打分结果,按照情感正负极性分为两类评价结果,并按照情感倾向程度的不同转化为直觉三角模糊数,从而实现模型的定量化。(3)基于前景理论的B2C消费者行为决策模型算例分析。对决策模型进行计算,主要应用了直觉三角模糊数的相关运算性质,求出各个属性下的前景值并得出综合前景值。通过与期望效用理论构建的决策行为模型进行比较,发现本文构建的基于前景理论的行为决策模型能更贴近实际情况。此外,通过对结果进行排序分析,讨论各个备选方案的优劣程度,得到各个备选方案的管理启示。研究结果表明:(1)总体上,京东商城综合前景得分最高,1号店次之,当当网最低。但分析累积综合得分发现,1号店在前期服务水平落后于当当网的情况下,通过后阶段高水平的服务提供,在总得分上超过了当当网,这对企业如何实现服务补救具有参考价值;(2网络口碑的聚类分析得到了13个评价指标,在这些指标中,京东商城在其中7个指标上前景值最优,1号店有4个前景值最优。京东商城虽然得分最高,但其售后服务阶段的得分却低于其余两者,导致总得分差距亦被缩小。而1号店最初得分较低,但通过出色的服务补救,最终得分排名第二。这充分说明,良好的服务补救策略对提升企业服务质量具有重要的作用。
[Abstract]:With the rapid development of e-commerce, online consumers show different behavior characteristics more and more. Such as: with the "double 11" activities, more and more consumers expect and accept, in the course of activities carried out, there are a large number of network consumers stay single brush, impulsive consumption (chop hand family), after a large number of return behavior. These behaviors in a certain extent reflects the network consumer behavior in the decision-making process of "limited rationality" and "irrational" characteristics. Therefore, according to the behavior characteristics of the current network of consumers, the purpose of this research is to reasonably correct modeling of the decision making process of the online shopping consumer "limited rationality" or "irrational" behavior, so as to provide feasible suggestions for enterprises to carry out operations management and service marketing management. The main contents of this research include: (1) the network shopping situation Consumer behavior research and decision modeling. To determine the stage of service process by using service blueprint, B2C e-commerce service process is divided into decision-making, access, transaction, payment, logistics and customer service of 6 sub process; the decision of customer behavior in the process of establishing the value of each function, weight function and function by introducing the prospect; hundreds of perfect intuitionistic fuzzy mathematical model; finally, through the definition of multiple attribute decision making problems, determine the structure of Jingdong, and the corresponding decision attribute and Dangdang Shop No. 1 set for alternatives. (2) quantitative decision model based on the word-of-mouth network. Based on the behavior decision model has been established on the mining IWOM quantitative representation decision model. Firstly, using the Java programming language, the classification of word of mouth network according to the service process, the realization of the SVM algorithm of network flow classification model based on reputation service points; The class is completed, call the Weka data analysis tools, internet word-of-mouth on intra class cluster analysis. The main use of EM algorithm to determine the number of clusters within class clustering (i.e. index number), and make sure each kind of neutron decision index weight assisted QFD research methods. After clustering, the sub process there are 13 decision ROST index; the further use of sentiment analysis tools emotional score on the network reputation, according to evaluation results, according to the polarity of the emotion is divided into two categories according to the evaluation results, and different degree of transformation sentiment intuitionistic triangular fuzzy number, so as to realize the quantitative model. (3) cases based on the B2C consumer behavior decision model prospect theory. The decision model is calculated, the main application of the correlation algorithm of triangle intuitionistic fuzzy number, calculated each attribute the prospect value and the result of comprehensive prospect value. With the expected utility theory Compare the decision behavior model, found the behavior decision model of prospect theory based on more practical built in this paper. In addition, sorting through analysis of the results, discuss the extent of the various alternatives, get the management implications of various options. The results show that: (1) overall, the Jingdong mall comprehensive prospect the highest score, 1 stores and the lowest dangdang.com. But analysis of cumulative score, shop No. 1 in the pre service level behind the dangdang.com case, by providing a high level of service after the stage, in the total score more than that of dangdang.com, enterprise how to implement service recovery has the reference value; (cluster analysis of 2 online word-of-mouth got 13 evaluation indexes in these indicators, the 7 indicators in the mall Jingdong to view the optimal value, shop No. 1 has 4 prospect value optimization. Although most score mall Jingdong High, but its customer service service stage is lower than that of the rest of the two scores, resulting in a total score gap was also reduced. The No. 1 shop was originally a low score, but through excellent service recovery, the final score ranked second. This shows that the good service recovery strategy plays an important role in promoting enterprise service quality.

【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F713.55

【参考文献】