Modeling And Analysis Of Consumers’ Willingness To Pay For Traceable Food: LCM And Computer Simulation

In recent years, food safety and quality issues are very prominent and they have raisedstrong concerns about it. Food traceability system is considered to be one of the main toolsof security risks through a reliable and continuous flow of information security in thesupply chain because it could monitor the flow and the food production process andidentify the source of the problem by tracing and implementation of the recall. In this paper,we research consumers’ willingness to pay for traceable food. For there is no large-scaledomestic food safety research data, we design menu based choice experiment by studyingwhat traceability information should be included in traceability food. Survey result wasused as input data. Since the type of survey data is categorical, the general method isdifficult to analyze it. To solve the above problems, latent class models and improvedk-modes clustering simulation were integrated to study consumer’s willingness to pay forthe traceability property in order to analyze consumer’ preference, expand consumerdemand for traceability food, better promote food traceability system and ensure foodsafety. The main work is as follows:(1) Through study the food traceability related literature and willingness to pay fortraceability property, we thought pork traceability supply chain should include breeding,slaughtering, processing and distribution of sales and government certified four properties.Questionnaire was designed include the above four traceability property in different pricelevel. Questionnaire design was tested by D-efficiency, test result showed design was good.Survey was conducted in Wuxi, Jiangsu province. Statistical analysis showed that thesurvey result was good.(2) We designed latent class model on the basis of utility theory to analyze consumers’willingness to pay for traceability information. Treat consumers’class as potential variablesand treat consumers’ choice as explicit variables. The menu choice experiment survey datawere used to analyze consumers’ behavior. The results show that consumer demand forfood traceability attributes showing different preferences, generally at a low level oftraceability food’consumption.(3) Through the study of clustering algorithms, we find the k-modes algorithm issuitable for analyzing menu choice experiment data which is characterized by discretecategorical data. For the complex process and not high classification accuracy, we combinethe latest research progress improve k-modes algorithm. Two factors, density and distance,were combined to select the initial cluster centers to simplify the process of clustering.Cluster modes was replaced by the pattern in which all property values are considered,thereby improving the clustering accuracy.(4) According to the survey questionnaire, a clustering model was established. Then theimproved k-modes clustering method was applied to analyze the menu methodquestionnaire data. We used the trend of CU and objective function to select the appropriatenumber of clusters. The results of this study show that consumers can be divided intoseveral groups which prefer different combinations of security information property. In addition, these groups have different ability to pay for traceability property. Food withdifferent combinations of traceability information could be provided to different customerclass thus expanding consumers’ demand for traceable food and finally improving foodsafety standards.