Sewage treatment fault diagnosis method based on stacking meta-learning strategy

Abstract

The invention discloses a sewage treatment fault diagnosis method based on a stacking meta-learning strategy. The method is to use the mean value method to complement the defective items of samples with incomplete attributes in the sewage data, and normalize them to the [0,1] interval Middle; set the optimal parameters of the number of hidden layer nodes, regularization coefficient and kernel width of the base classifier; use the processed training samples to perform 3-fold cross-validation on the base classifier to obtain the original output of the base classifier; set the base classifier The original output of the algorithm is transformed into a probabilistic output, and a meta-training set is constructed; the meta-classifier is trained using the meta-training set to obtain the final classification decision model. The invention integrates different base classifiers through the meta-learning strategy, improves the diversity and stability of the algorithm, and further improves the overall performance of fault diagnosis in the process of sewage treatment.

Description

technical field [0001] The invention relates to the technical field of sewage treatment fault diagnosis, in particular to a sewage treatment fault diagnosis method based on a stacking meta-learning strategy. Background technique [0002] Wastewater treatment is a complex, multivariable biochemical process. The failure of sewage treatment plants can easily lead to a series of serious sewage pollution problems. The fault diagnosis of sewage treatment process can be transformed into a classification problem of pattern recognition. The sewage data is composed of the data collected regularly by the sewage plant and its current working status, and the sewage data within a period of time constitutes the sewage data set. Due to the low frequency of faults in qualified sewage treatment plants, the data in the fault state in the collected sewage data set is often far less than the data in the normal state. This results in a highly unbalanced distribution of the sewage data set, tha...

AI Technical Summary

Problems solved by technology

This leads to a highly unbalanced distribution of the sewage data set, that is, the fault diagnosis of the sewage treatment process is a classification problem with unbalanced data.

[0003]Traditional learning algorithms often optimize parameters based on the overall accuracy rate, which tends to make the classification results more biased towards the majority class

Images