A Risk Assessment Optimization Model Based on TOPSIS Algorithm
Yunkai Liu · 2022
With the development of information technology and data processing technology, mining data to reasonably monitor and avoid production risks in real time can provide an important guarantee for enterprise production. Combining the current mainstream research - mathematical modeling analysis and machine learning analysis, this paper puts forward the production risk monitoring and evaluation method based on the detector data of a production enterprise. First, the Median Absolute Deviation Algorithm is used to detect the abnormal value of the real-time data recorded by the sensor. On this basis, the autocorrelation coefficient and Spearman correlation coefficient are used to calculate, and the characteristics of 'Contingency', 'Persistence' and 'Linkage' are extracted as distinguishing the abnormal risk, and 28 groups of detectors are obtained; Finally, 'Deviation Degree', 'Change Rate', 'Abnormal Linkage' and 'Abnormal Persistence' are extracted as evaluation indicators, and a gray correlation degree evaluation model is established to evaluate the risk degree at each time, so as to calculate the risk level at each time. The experimental results show that there are high probability production risks in about 8:42PM and 10:50PM, which need to be paid more attention to. This paper provides a new idea for production monitoring in actual enterprise operation.