Active learning for spectroscopic data regression

In this work, we introduce an active learning approach for the estimation of chemical concentrations from spectroscopic data. Its main objective is to opportunely collect training samples in such a way as to minimize the error of the regression process while minimizing the number of training samples used, and thus to reduce the costs related to training sample collection. In particular, we propose two different active learning strategies developed for regression approaches based on partial least squares regression, ridge regression, kernel ridge regression, and support vector regression. The first strategy uses a pool of regressors in order to select the samples with the greatest disagreements among the different regressors of the pool, while the second one is based on adding samples that are distant from the current training samples in the feature space. For support vector regression, a specific strategy based on the selection of the samples distant from the support vectors is proposed. Experimental results on three different real data sets are reported and discussed. Copyright © 2012 John Wiley & Sons, Ltd.     

Assigning environmental variables to observed biological changes

A method for assigning environmental variables to observed biological changes in benthic communities is proposed. The approach requires biological and environmental sampling at the same sites. Additionally, a biological gradient or trend such as a change in observed species or a significant change in their relative abundances is necessary in order to connect the biological observations to the environmental measurements. Whether there is a statistical significant correspondence between the environmental measurements and the biological changes is tested after quantifying the biological changes by using the community disturbance index (CDI). Finally, the environmental variables that are most strongly associated with the biological changes are identified, and it is proposed that these are strong candidates as the pollutants responsible for the biological changes observed. However, this cannot be confirmed using the monitored data only. The approach is tested on data collected in monitoring surveys at the Ekofisk oil field in the North Sea. The results indicate the method is feasible for assigning environmental variables to observed biological changes.     

基于可见/近红外光谱分析技术的猪肉肉糜品质检测研究

以225个猪肉肉糜样本为研究对象,利用可见/近红外光谱分析技术对猪肉肉糜主要品质指标的的快速检测进行了研究.光谱经小波去噪后,采用偏最小二乘法和支持向量机定量分析方法分别建立了肉糜中肌内脂肪、蛋白质和水分含量的可见/近红外光谱预测模型.其中,肌内脂肪的支持向量机定量预测模型最优,校正相关系数rcal和预测相关系数rva...     

基于二次迭代Monte Carlo的羊肉硬度定量检测研究

随着居民生活水平的提高和对健康饮食结构的重视,羊肉作为一种高蛋白且低脂肪和胆固醇的畜肉,需求量逐年上涨.根据国家统计局统计,2012年—2019年我国畜肉产业中羊肉产量占比从6.27％ 上升到9.02％.研究提出了一种基于二次迭代Monte Carlo(MC)算法剔除异常样本的羊肉硬度定量检测PLSR模型.采用Gaia...     

铁皮石斛多糖含量的高光谱反演研究

铁皮石斛是我国传统的中草药,多糖含量是衡量其品质优劣的一个重要指标。主要研究铁皮石斛多糖含量与光谱间的关系,旨在寻找铁皮石斛多糖含量的无损检测方法。研究方法:实验材料选自不同处理铁皮石斛组培苗和不同生长阶段的驯化苗。其中36个样本作为建模样本建立模型,11个样本作为检验样本对模型进行检验。分别采用偏最小二乘回归法和因子分析法建立样本光谱与多糖含量间的关系。结果表明:(1)光谱一阶微分与多糖含量的相关性优于光谱反射率,其中对多糖含量较敏感的波段主要集中在可见光区域。(2)偏最小二乘回归法的模型决定系数最高,但其预测能力较差。因子分析法模型具有较好预测能力,基于光谱反射率因子分析模型和基于光谱一阶微分因子分析模型的相对分析误差分别为2.269和2.305。