A big data quality evaluation method based on group heterogeneity rationality perception information fusion

Abstract

Big data is complex, and it is difficult to evaluate big data quality with only a few evaluation subjects. This paper designs a universal evaluation method for big data quality based on the idea of group heterogeneity rationality perception information fusion. Group evaluation information is usually large-scale and heterogeneous. The linear programming technique for multidimensional analysis of preference (LINMAP) model has the performance of large-scale evaluation information processing. Because the existing LINMAP models cannot effectively process group evaluation information with heterogeneous rationality, the LINMAP model of heterogeneous rationality (HR-LINMAP) based on group evaluation information is proposed. The research results show: under certain conditions, the LINMAP model of bounded rationality (BR-LINMAP) can be transformed into the LINMAP model of perfect rationality (PR-LINMAP), and the HR-LINMAP is the general form of PR-LINMAP and BR-LINMAP. The scientific research big data resources with mature trading markets are selected as the evaluation objects, and HR-LINMAP is used to analyze the big data quality evaluation cases. The empirical analysis shows that the HR-LINMAP model is feasible and effective, which provides a new idea for big data quality evaluation.