Data Driven Quantitative Trust Model for the Internet of Agricultural Things

With frequent food safety incidents, it becomes urgent and important to design an intuitively quantitative trust model to describe the trustworthiness of foods delivered in supply chains. However, current existing models are usually too subjective, because they heavily depend on experts experiences to model the trust and set relevant parameters. Fortunately, the Internet of Agricultural Things may offer a big volume of business data, including product information and delivery information etc., via its pervasive sensing. These data motivate us to design a data driven quantitative model to evaluate the trust of sensed products in supply chains. The proposed trust model leverages a Bayesian network, where almost all parameters are set by the data rather than experts experiences, to evaluate the trust value of a target product. Finally, a case of pork product is used to show the effectiveness of our trust model. Based on the comparison with other models, our model is promising to reduce the subjectivity and time-delay of the trust evaluation.