Comprehensive Evaluation Modeling and Analysis Based on ELM Integrated AHP and PCA: Application to Food Safety

Abstract

As an effective investigation, quality evaluation can summarize the various aspects of the corresponding evaluation objects. To reflect the true situation of the evaluation object more comprehensively and accurately, the focus of the evaluation process is how to effectively deal with the data of each evaluation indicator. To this end, this paper proposes a quality evaluation and analysis model, which is based on extreme learning machine (ELM) integrated principal component analysis (PCA) and analytic hierarchy process (AHP). The proposed model makes full use of the systematic and simplicity of AHP to assign weights to several important indicators, and uses PCA to eliminate the relevant impact of evaluation indicators, reducing the workload of indicator selections. Then the comprehensive quality score of the evaluation object is calculated according to the comprehensive evaluation function. At last, the generated new sample data set constitutes the training set of ELM, and finally the comprehensive quality evaluation model is obtained. To evaluate the performance of the proposed model, the proposed model is applied to the food safety data processing. Compared with the traditional food risk analysis model, the proposed model can get a more comprehensive result, which can enable decision makers to grasp the food quality information more accurately and comprehensively, and make corresponding feedback timely.