基于区间直觉模糊数的多属性决策模型在顾客需求偏好中的应用

Application of Multi-Criteria Decision Model Based on Interval Intuitionistic Fuzzy Number in Customer Demand Preference

摘要: 随着电子商务的快速发展，消费者的评论在顾客的需求偏好中的影响越来越大。但是，客户的评论中往往蕴含着大量的不确定信息，很难使用精准数据去刻画决策数据。而且现有研究中的属性权重确定大多是由专家直接确定，没有考虑到消费者的个人偏好。针对上述问题，本文提出了一种基于区间直觉模糊数的多属性决策模型来解决顾客的需求偏好问题。首先，引入区间直觉模糊的思想，将其转化为多准则决策(MCDM)问题，并以区间直觉模糊数的形式来刻画其中包含的决策信息；其次，使用BWM算法来计算指标权重，科学、准确地反映出顾客对于每个指标的个人偏好信息；继而，基于比率分析的区间直觉模糊多重目标优化(IVIF-MULTIMOORA)对产品进行排名，使用三种方法来建立鲁棒的决策。最后，通过拥有不同需求偏好的两个顾客进行实例分析，分别确定两名顾客不同的备选产品排序，找出最适合他们的产品，并且通过对比试验和灵敏度分析说明了本文的有效性。

Abstract: With the rapid development of e-commerce, consumers’ reviews have more and more influence on customers’ demand preferences. However, customer reviews often contain a lot of uncertain information, and it is difficult to describe decision data with accurate data. In addition, most of the existing studies on determining criteria weight are directly determined by experts, without considering the personal preferences of consumers. Aiming at the above problems, this paper proposes a multi-criteria decision model based on interval intuitionistic fuzzy number to solve the problem of customer demand preference. Firstly, the concept of interval intuitionistic fuzzy is introduced, which is transformed into multi-criteria decision making (MCDM) problem, and the decision information contained in it is described with the form of interval intuitionistic fuzzy number. Secondly, BWM algorithm is used to calculate the attribute weight, which reflects the customer’s personal preference information for each attribute scientifically and accurately. Then, the products were ranked using the interval Intuitionistic Fuzzy Multiple Objective Optimization based on a ratio analysis plus a full multiplicative form (IVIF-MULTIMOORA) method that uses three methods to establish a robust decision. Finally, through the case analysis of two customers with different demand preferences, the order of alternative products of two customers is determined respectively, find out the most suitable products. And the effectiveness of this paper is illustrated through comparative test and sensitivity analysis.