消费者对可追溯食品支付意愿建模及分析：潜类别模型及计算机模拟

发布时间：2019-03-12 10:15

【摘要】：近年来，国内食品安全质量问题十分突出，引发了人们对食品安全的强烈关注。食品可追溯体系通过在供应链上形成可靠且连续的安全信息流，能够监控食品生产过程与流向且通过追溯来识别问题源头和实施召回，被认为是有效消除信息不对称，从根本上预防食品安全风险的主要工具之一。本文对消费者对可追溯属性支付意愿展开研究。针对国内没有大规模的食品安全调研数据，本文通过研究食品可追溯性应包含的可追溯信息与层次，运用菜单选择实验法设计问卷进行调研，将调研结果作为研究分析的数据输入。由于菜单法数据是离散的分类属性数据的特点，一般方法难以应用。针对以上问题本文综合运用经济学的潜类别模型和适用于分类属性数据的k-modes聚类算法研究消费者对猪肉可追溯属性的支付意愿从而分析消费者的群体性偏好，以扩大消费者对可追溯食品的需求，更好的推广食品可追溯体系，保障食品安全。本文主要工作如下： (1)通过对国内外食品可追溯体系及支付意愿相关文献研究与实地调研，设计猪肉供应链体系的可追溯信息应该包括养殖、屠宰加工和配送销售及政府认证四个属性。对上述四个可追溯属性分别设置不同的价格层次设计菜单选择实验法问卷，以D-efficiency检验显示问卷设计优良。在江苏省无锡市实地问卷调研收集数据,问卷结果统计分析显示调查结果良好。 (2)根据消费者效用理论针对消费者对可追溯支付意愿建立潜类别模型，以消费者的类别为潜在变量，以消费者的选择作为外显变量，运用菜单选择实验法调查数据对消费者行为分析。结果表明消费者对食品可追溯属性的需求呈现出不同的偏好，普遍属于低水平可追溯属性消费群体。 (3)通过对聚类相关算法研究，针对菜单选择实验法数据是离散的分类属性数据的特点，找到适合聚类分析的k-modes算法。针对k-modes算法存在的聚类过程复杂，分类精度不高的问题结合最新研究进展综合改进k-modes聚类，通过结合密度和距离两个因素选取初始聚类中心从而简化聚类过程，以考虑可追溯属性的所有属性值的模式代替k-modes聚类算法的modes，从而提高聚类分类精确性。 (4)根据调查问卷建立聚类分析模型，将改进k-modes聚类方法应用于菜单选择实验法问卷结果分析，，以CU和目标函数走向来选取合适的聚类类别数目。研究结果表明，消费者可分为多个对可追溯安全信息偏好不同的群体，这些群体支付能力也不同。可针对不同的群体提供不同的可追溯属性组合的猪肉以扩大消费者对可追溯食品的需求，提高食品安全保障水平。
[Abstract]:In recent years, the domestic food safety quality problem is very prominent, has aroused the people to the food safety strong concern. The food traceability system is considered to be an effective way to eliminate information asymmetry by creating a reliable and continuous information flow across the supply chain, monitoring the food production process and direction, and identifying the source of the problem and implementing the recall through traceability. One of the main tools for fundamental prevention of food safety risks. This paper studies consumers' willingness to pay for traceability attributes. In view of the lack of large-scale food safety research data in China, this paper studies the traceability information and levels of food traceability, and designs a questionnaire by menu selection experiment. The results of the research are input as the data of the research and analysis. Because menu method data is the characteristic of discrete classification attribute data, the general method is difficult to apply. In order to solve the above problems, this paper studies consumers' willingness to pay for pork traceability attributes by using the latent category model of economics and k-modes clustering algorithm, which is suitable for classification attribute data, so as to analyze consumers' group preference. In order to expand the consumer demand for traceability food, better promotion of food traceability system, food safety. The main work of this paper is as follows: (1) through the literature research and field investigation of domestic and foreign food traceability systems and willingness to pay, the design of pork supply chain system traceability information should include breeding, Slaughtering processing and distribution sales and government certification four attributes. The above four traceability attributes are set up with different price level design menu to choose the experimental questionnaire. The D-efficiency test shows that the questionnaire design is excellent. The data were collected in Wuxi City, Jiangsu Province, and the results of statistical analysis showed that the results were good. (2) according to the consumer utility theory, a latent class model is established for consumers' willingness to pay retroactively, taking the consumer's category as the potential variable and the consumer's choice as the explicit variable. The method of menu selection experiment was used to analyze consumer behavior. The results show that consumers' demand for food traceability attribute presents different preferences, and generally belongs to the low-level traceability consumer group. (3) according to the characteristic that menu selection experiment data is discrete classification attribute data, the k-modes algorithm suitable for clustering analysis is found through the research of clustering correlation algorithm. In view of the complex clustering process and the low classification accuracy of k-modes algorithm, the clustering process is simplified by selecting the initial clustering center by combining the density and distance factors, combined with the latest research progress, and the k-modes clustering is improved synthetically by combining the two factors of density and distance. The modes, of k-modes clustering algorithm is replaced by the pattern considering all attribute values of traceability attributes, so as to improve the accuracy of clustering classification. (4) the cluster analysis model is established according to the questionnaire, and the improved k-modes clustering method is applied to the analysis of the results of the menu selection experiment. The CU and objective functions are used to select the appropriate number of cluster categories. The results show that consumers can be divided into several groups with different preference for traceability security information, and these groups have different ability to pay. Different groups of pork can be provided with different traceability properties to expand consumers' demand for traceability food and improve food safety.
【学位授予单位】：江南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TS201.6;TP311.13

【参考文献】