Auxiliary variable selection method considering causal effect in industrial soft measurement

Abstract

The invention discloses an auxiliary variable selection method considering a causal effect in industrial soft measurement, and the method comprises the steps: obtaining a discretized candidate variable set and a dominant variable data sample through preprocessing based on a historical data set of a soft measurement modeling object; determining the causal effect of each candidate variable and the dominant variable through the mutual information of each candidate variable and the dominant variable; adopting a variable screening algorithm based on the causal effect and saving all candidate variables of which the causal effect is not zero as an auxiliary variable set. The causal effect between the candidate auxiliary variable and the main variable is considered, the method does not depend on any model, a stop threshold does not need to be set, the candidate variable combination with the causal effect not being zero is automatically selected as the auxiliary variable set, and useful reference is provided for industrial soft measurement modeling.

Description

technical field [0001] The invention relates to a technology in the field of information processing, in particular to an auxiliary variable selection method considering causal effects in industrial soft sensors. Background technique [0002] Soft measurement uses difficult-to-measure key performance indicators as the output (leading variables), and selects related and easy-to-measure variables as inputs (auxiliary variables), and constructs a certain mathematical relationship to realize the prediction and estimation of the leading variables. It is used to evaluate indicators such as product quality, production efficiency, energy consumption, and pollutant discharge. Obviously, how to select the appropriate auxiliary variables has become the primary problem of industrial soft sensor, which directly determines the complexity, prediction accuracy and application reliability of the soft sensor model. Existing auxiliary variable selection methods are generally based on: domain e...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problems that the existing auxiliary variable selection method depends on the threshold or model and has poor interpretability, the present invention proposes an auxiliary variable selection method considering causal effect in industrial soft sensor, by considering the causal effect between the candidate auxiliary variable and the leading variable , does not depend on any model, does not need to set a stop threshold, and automatically selects candidate variable combinations with non-zero causal effects as auxiliary variable sets, providing a useful reference for industrial soft sensor modeling

Images