The Design of University Coordination Utility Management and Online Repair Platform Based on Multivariate Statistical Analysis with Random Matrix

Abstract

In this paper, the random matrix of multivariate statistical analysis is used to conduct in-depth research and analysis of the university coordination utility management and online repair platform. Considering that the chunking of variables based on mechanistic knowledge is not easy to achieve, firstly, the maximum correlation and minimum redundancy algorithm is used to portray the correlation more accurately between process variables and remove the redundancy between variables to provide the optimal variable input for the base model. The multivariate mean control chart was used to calculate the offset between the data of each test group of the contact network and the overall mean and standard values of the contact network parameters under different correlations among the contact network parameters. Based on the daily work research and process document sampling of the university coordination utilities management department, the requirement analysis and design of the target system were completed, and a university coordination utility management system based on BS architecture was developed. Student information is lost, data statistics are wrong, etc., so that the business work of other departments of the school cannot be carried out smoothly. The whole platform can be divided into several submodules according to the functions: super administrator module, administrator module, staff module, and user module, and the detailed design scheme of each module is described in detail. At the same time, the logistic regression model is trained using the collected data sets, and the training scheme of the model is designed. The mathematical model of logistic regression and the related algorithm are used to decide whether to purchase maintenance equipment at this stage and the quantity of purchase. Finally, a new monitoring index is proposed to monitor the process status. MNPE-GMM not only maintains most of the local structural information of the window dataset in the feature subspace but also reduces the computational complexity of GMM in the fault detection process. The MNPE-GMM method can effectively improve the fault detection rate of multimodal intermittent processes by introducing new statistics.