Early warning modeling and analysis based on analytic hierarchy process integrated extreme learning machine (AHP-ELM): Application to food safety

Abstract

Since the actual food safety monitoring data have characteristics of high-dimension, complexity, discreteness and nonlinear properties, it is difficult to accurately predict the risk of actual food inspection process. Therefore, this paper proposes a predictive modeling approach based on analytic hierarchy process (AHP) integrated extreme learning machine (ELM) (AHP-ELM). The proposed approach utilizes the AHP model to obtain the effective process characteristic information (PCIs). Compared with the analytic hierarchy process (AHP) integrated traditional artificial neural network (ANN) approach, the AHP-ELM prediction model is effectively verified by executing a linear comparison between all PCIs and the effective PCIs through daily inspection data source from the supervision and inspection department repository of China quality supervision system. Finally, the PCIs and the prediction value are obtained to provide more reliable food information and identification of potentially emerging food safety issues. The proposed method is applied to the food safety early warning and monitoring system in China. The result shows that the proposed model is effective and feasible in processing the complex food inspection data. Meanwhile, it can help to improve the quality of food products, ensure food safety and reduce the risk of food safety.