论文部分内容阅读
目的采用可见/近红外光谱技术,结合岭回归偏最小二乘对猪肉新鲜度进行定量分析。方法利用自行搭建的可见/近红外光谱检测系统,采集62个猪肉样品表面380~900 nm范围内的反射光谱数据,进行标准正态变量变换(standard normal variable transform,SNVT)预处理后,建立偏最小二乘(partial least square regression,PLSR)模型。利用模拟退火算法(simulated annealing,SA)和粒子群算法(particle swarm optimization,PSO)进行岭参数寻优,建立猪肉挥发性盐基氮(total volatile basic nitrogen,TVB-N)的岭回归模型。结果所建模型的相关系数和误差分别为0.9819、1.2785 mg/100 g和0.9781、1.4628 mg/100 g。结论所建立的模型取得了较好的结果,利用岭回归偏最小二乘实现了对最小二乘估计的改良,更加验证了可见近红外光谱技术对猪肉新鲜度进行定量分析的巨大应用潜力。
OBJECTIVE To quantitatively analyze pork freshness by visible / near infrared spectroscopy combined with Ridge regression partial least squares. Methods The visible / near-infrared (NIRS) detection system was used to collect the reflectance spectra of the surface of 62 pork samples in the range of 380-900 nm. After the standard normal variable transform (SNVT) pretreatment was performed, Partial least square regression (PLSR) model. The ridge parameters were optimized by simulated annealing (SA) and particle swarm optimization (PSO), and ridge regression model of pork volatile basic nitrogen (TVB-N) was established. Results The correlation coefficient and error of the model were 0.9819, 1.2785 mg / 100 g and 0.9781, 1.4628 mg / 100 g respectively. Conclusion The established model has achieved good results. Ridge regression partial least squares method is used to improve the least squares estimation, which further verifies the great potential of visible near - infrared spectroscopy for the quantitative analysis of pork freshness.