论文部分内容阅读
将近红外光谱技术和化学计量学相结合分析大黄鱼新鲜度指标K值。大黄鱼原始光谱经多元散色校正,联合区间偏最小二乘法(siPLS)优化建模区域,建立新鲜度分析模型,并与经典偏最小二乘法模型和间隔偏最小二乘法模型相比较。结果表明,当采用联合区间偏最小二乘法将全光谱划分为16个区间,2个子区间联合(2,3)时,建立的siPLS模型预测效果最好,其交互验证均方根误差(RMSECV)和预测均方根误差(RMSEP)分别为3.63和3.49,校正集和预测集的相关系数分别为0.98059和0.91287。利用联合区间偏最小二乘算法,可有效地减少建模所用变量数,实现大黄鱼新鲜度的快速检测。
Near infrared spectroscopy and chemometrics were combined to analyze the freshness index K of large yellow croaker. The original spectra of the large yellow croaker were optimized by multivariate random color correction and combined with partial least squares (siPLS). The freshness analysis model was established and compared with the classical partial least square method and interval partial least squares method. The results show that the siPLS model is the best one to predict when the whole spectrum is divided into 16 intervals and 2 subintervals (2, 3) using the joint interval partial least squares method. The root mean square error of validation (RMSECV) And the root mean square error of prediction (RMSEP) were 3.63 and 3.49, respectively. The correlation coefficients between the calibration set and the prediction set were 0.98059 and 0.91287, respectively. The joint interval partial least-squares algorithm can effectively reduce the number of variables used in modeling and achieve rapid detection of freshness of large yellow croaker.