论文部分内容阅读
应用R语言rioja软件包的加权平均(Weighted Averaging,WA)和加权平均偏最小二乘(Weighted AveragingPartial Least Squares,WA-PLS)模型建立了长白山区泥炭藓泥炭地有壳变形虫与水位埋深(depth to water table,DWT)、pH和泥炭湿度的转换函数,为古环境定量重建奠定了基础,也提供了rioja软件包应用的实例和参考。结果表明水位埋深以WA-PLS模型最佳(预测均方根误差RMSEP为7.39 cm,R2=0.74);对于pH和泥炭湿度,WA-PLS第一分量和WA.inv都产生了最小的RMSEP和较高的R2值。pH的RMSEP为0.18,R2为0.72。泥炭湿度的RMSEP为1.95%,R2为0.62。如果泥炭剖面的有壳变形虫种类组成与本研究的训练样本集相同,水位埋深、pH和泥炭湿度可以分别以±7.39 cm、±0.18和±1.95%的平均误差进行重建。
A Weaned Averaging (WA) and Weighted Averaging Partially Least Squares (WA-PLS) model of the rioja software package was used to establish the crustacean shelled pests in Changbai Mountain peat moss and the water depth depth to water table (DWT), the conversion function between pH and peat moisture, which laid the foundation for the quantitative reconstruction of paleoenvironment. It also provided examples and references for the application of rioja software package. The results showed that the WA-PLS model was the best for the water depth (RMSEP was 7.39 cm and R2 = 0.74). For the pH and peat moisture, the first component of WA-PLS and WA.inv all produced the smallest RMSEP And higher R2 value. The RMSEP of pH is 0.18 and R2 is 0.72. RMSEP for peat moisture is 1.95% and R2 is 0.62. If the crustacean species in the peat profile is the same as the training sample set in this study, the water depth, pH and peat moisture can be reconstructed with mean errors of ± 7.39 cm, ± 0.18 and ± 1.95%, respectively.