【摘要】 通过对选取的离散小波变换后的小波细节系数的重构光谱进行iPLS建模,从中获得了较好的预测结果。实验表明,合理选择小波细节系数可减少非主要信息的干扰,并以此来提高成分分析精度的方法是切实可行的,采用iPLS建立的玉米蛋白质模型,不仅能有效地减少建模所用的变量数,缩短运算时间,而且预测能力和精度均得到提高,在实际应用中有一定的参考价值。更多还原

【Abstract】 It establish iPLS modeling for reconstructing spectral of wavelet detail coefficients after discrete wavelet transform,and better forecasts is obtained.The experiments show that reasonable choice of wavelet detail coefficients can reduce interference of non-main information,and it is a feasible method to improve the analysis precision.The corn protein model established by iPLS can decrease the variable number of modeling effectively and shorten the operation time,and can also improve the prediction ability and precision.It has certain reference value in the practical application.更多还原