Self-adaptive parameter soft measuring method on basis of semi-supervised local linear regression

Abstract

The invention provides a self-adaptive parameter soft measuring method on the basis of the semi-supervised local linear regression, which comprises the following steps of: firstly, on the theoretical basis of a local linear regression scatter point flatting method, transforming the local linear regression scatter point flatting method into a semi-supervised learning method by a method of introducing unlabeled samples into a target equation of the unlabeled samples; then by utilizing a method of calculating a labeled sample set to estimate a slope, carrying out self-adaptive estimation to obtain an optimal gaussian kernel width parameter of a current labeled sample set; and finally, implementing parameter selection of the semi-supervised local linear regression learning method by utilizing a self-adaptive parameter selecting method and implementing the online update of a soft measuring model on the basis of a sliding time window mode. The invention sufficiently utilizes the value of the unlabeled samples. Aiming at the characteristics of the semi-supervised local linear regression, the influence caused by the labeled sample measuring error can be effectively eliminated and the prediction accuracy is improved. The soft measuring model is updated in real time by applying the sliding time window method, so that the model can well adapt to the change of input data.

Description

technical field [0001] The invention relates to a method in the field of soft sensor technology, in particular to an adaptive parameter soft sensor method based on semi-supervised local linear regression. Background technique [0002] Soft measuring instrument refers to a method of using computer modeling technology to measure industrial process control variables. This method is different from traditional sensor online measurement and manual offline measurement. It has the advantages of low investment cost, simple maintenance and high reliability. advantage. Currently existing soft sensor methods are generally based on supervised learning methods, which only use labeled samples for model training, thus greatly wasting those unseen samples that can be easily obtained in practice and actually reflect the operation of industrial processes. The value of the labeled sample. If these unlabeled samples can be used in the modeling of soft sensor models, it should be possible to im...

AI Technical Summary

Problems solved by technology

[0005] The present invention aims at the problem that the unused unmarked samples and the model cannot be effectively updated in the traditional soft sensor method, and provides an online soft sensor method based on a semi-supervised local linear regression algorithm. The method uses the local linear regression scatter point smoothing method as the Theoretical basis, by introducing unlabeled samples into its manifold regularization function, it is changed into a semi-supervised regression method, which is used to perform soft sensor modeling on labeled samples and unlabeled samples for prediction of control variables or Measurement

Images