Pork quality identification based on principal component analysis and improved support vector machine

Abstract

Objective: In order to eliminate the large amount of redundant information in near-infrared spectroscopy and to improve the accuracy of pork quality identification, and to establish a method for rapid identification of pork quality. Methods: Principal component analysis was used to reduce the dimensionality of pork spectrum data and the characteristic wavelengths of pork spectrum were selected. The parameters of the support vector machine （SVM） model were optimized by the salp swarm algorithm. Pork quality recognition model was proposed based on an improved support vector machine optimized by salp swarm algorithm. Results: compared with grey wolf optimization algorithm improved SVM （GWO-SVM）, grid search algorithm improved SVM （Grid-SVM）, particle swarm optimization algorithm improved SVM （PSO-SVM） and SVM, the pork quality recognition model based on SSA-SVM had the highest precision. Conclusion: Pork quality identification model based on PCA and SVM optimized by salp swarm algorithm can effectively improve the accuracy of pork quality identification.

Keywords

• near-infrared spectroscopy
• support vector machine
• salp swarm algorithm
• principal component analysis
• particle swarm optimization algorithm
• pork quality