Method for constructing cocoon vacuum water bath process design reelability prediction model based on LSSA-LSSVM

Abstract

The invention discloses a silkworm cocoon vacuum water bath process design reelability prediction model construction method based on LSSA-LSSVM, and the method specifically comprises the following steps: obtaining a silkworm cocoon vacuum water bath process reelability data set, carrying out the normalization processing, and dividing the data set into a training sample set and a test sample set; setting an LSSVM initial parameter and establishing an LSSVM initial model, and performing training prediction on the training sample set by using the LSSVM initial model; performing parameter optimization on the prediction result by adopting an improved sparrow search algorithm LSSA to obtain an optimal hyper-parameter group (gamma *, sigma 2 *), so as to establish an LSSA-LSSVM-based silkworm cocoon vacuum water bath process design reelability prediction model; the test sample set is subjected to prediction test, and the prediction effect of the cocoon vacuum water bath process design reelability prediction model of the LSSA-LSSVM on the reelability is measured. The method has the advantages of being high in parameter optimization capacity, short in prediction time, high in prediction precision and the like, and reference is provided for relevant process designers to adjust and optimize the silkworm cocoon vacuum water bath process parameters.

Description

technical field [0001] The invention belongs to the technical field of raw silk production and processing, and in particular relates to a method for constructing a prediction model for unwinding rate of cocoon vacuum water bath process design based on LSSA-LSSVM. Background technique [0002] Cocoon cooking is a key process in the silk reeling process. The quality of cocoon cooking not only affects the efficiency of subsequent processing processes, but also directly affects raw silk quality indicators such as unwinding rate, cleanliness, cleanliness, and raw silk linear density. [0003] At present, silk reeling enterprises and fiber inspection institutions mostly use cocoon vacuum water bath process trial cooking method to verify whether the unwinding effect meets the actual production requirements, that is, by adjusting the processes of vacuum infiltration, low-temperature infiltration, steam infiltration, high-temperature water boiling and cooling to produce cocoons Combi...

AI Technical Summary

Problems solved by technology

[0004] In recent years, artificial intelligence methods have developed vigorously. Least Squares Support Vector Machine (LSSVM) has been applied in the textile field due to its strong nonlinear modeling ability, strong generalization ability, and short training time. However, the prediction performance of LSSVM Affected by hyperparameters, the prediction accuracy is not high

For this reason, some researchers use particle swarm optimization (PSO) to optimize LSSVM parameters. Although they can improve the performance of LSSVM parameters, these algorithms have the disadvantage of slow convergence speed and easy to fall into local optimum.

Images