Local label propagation-based instant learning semi-supervised soft measurement modeling method

Abstract

The invention relates to an instant learning semi-supervised soft measurement modeling method based on local label propagation, and the method comprises the steps: fully extracting information in an unlabeled sample through a local label propagation algorithm, building an online model related to query data through an overall optimization instant learning algorithm, and adding the online model as a constraint item into the local label propagation algorithm. For collected query data, firstly, a local model is established through an overall optimization instant learning algorithm according to marked historical data, then, similar samples of the query data are selected from all historical samples, and finally, an output predicted value of the query data is calculated through a local label propagation algorithm based on local model constraint. According to the method, the nonlinear, time-varying and multi-collinearity problems of the industrial process can be well solved, a large amount of unmarked historical data can be effectively utilized, and the utilization rate of historical samples and the prediction precision of a soft measurement model are improved.

Description

technical field [0001] The invention belongs to the technical field of industrial process detection, relates to the soft measurement technology of the industrial process, and in particular relates to a real-time learning semi-supervised soft measurement modeling method based on local label propagation. Background technique [0002] In the modern industrial production process, many important quality variables (such as oil viscosity, components, etc.) are difficult to measure in real time, which has a great impact on chemical process control and optimization. Due to the difficulties of on-site sampling, high cost of analytical instruments and lag in analysis time in the chemical production process, in the actual production process, it is often difficult to use online analytical instruments and offline assays to measure quality variables in real time, and it is impossible to form quality variables. Closed-loop control. Therefore, how to obtain quality variables in real time be...

AI Technical Summary

Problems solved by technology

[0006] The present invention aims at the above-mentioned problems that the existing just-in-time learning technology cannot use unlabeled historical samples, etc., and provides a semi-supervised soft sensor modeling method for just-in-time learning based on local label propagation (Just in time learning algorithm based on local label propagation, LLPJITL ), extending the real-time learning method to the semi-supervised field, which can efficiently extract the information contained in unlabeled samples, and improve the model optimization efficiency and prediction accuracy

Images