相思树种子的质量评价模型

投影寻踪技术为一种新兴的非线性分析方法,但存在计算过程复杂、编程实现困难等缺陷,鉴于此,我们提出采用模拟退火法直接优化投影寻踪技术的投影函数和投影方向以克服目前投影寻踪技术的种种缺点.在此基础上,提出基于模拟退火法的改进的投影寻踪的相思树种子质量评价模型,利用该模型可把各相思树种批测定指标综合成一维投影值,投影值越大,表示该种批质量越好,根据投影值的大小可对相思树种批样本集进行合理评价.实例质量评价结果表明,直接由样本数据训练的改进投影寻踪林木种子质量评价模型用于相思树种子质量评价简单可行,具有较强的适用性和应用性,可操作性强.基于模拟退火法的改进的投影寻踪法可广泛应用于种子质量评价、群落分类及环境质量评价等研究中各类非线性、高维数据的分级与评价.表1参11

Quality Assessment Model of Acacia spp. Seeds

The projection pursuit technique is a new method in nonlinear analysis, but it has some shortcomings such as complex calculating process and difficulty in programming. The simulated annealing method (SAM) was used to optimize the projection function and direction of projection pursuit technique directly, so as to overcome its shortcomings. In this paper, the quality assessment model (QAM) of Acacia spp. seeds was established by the modified projection pursuit method based on SAM, which was utilized to integrate the measured indices of Acacia spp. seeds into one-dimensional projection value. The samplings of Acacia spp. seeds can be classified into three types reasonably according to the projection value, which indicated that high projection value could generally reflect good quality of the seeds. The classification of examples implied that QAM trained by samples data directly was simple, feasible, applicable and maneuverable. The modified projection pursuit method based on SAM could be applied to classify and estimate all kinds of non-linear and multi-dimensional data in seed grading, community classification and environmental quality assessment. Tab 1, Ref 11