基于投影寻踪的滨海盐碱地改良综合效应评价

Evaluation of the combined effect of coastal saline-alkali soil reclaim based on the theory of projection pursuit

为综合评价滨海盐碱地改良综合效应,从经济效益、土壤质量、生态环境3个方面建立滨海盐碱地暗管改碱效应评价指标体系。应用基于粒子群算法的投影寻踪方法,建立滨海盐碱地暗管改碱效应等级评价模型,详细介绍了该模型的构建过程。利用该模型将暗管改碱效应等级的多维分级指标综合成一维投影值,根据投影值与评价等级经验值的函数关系可对暗管改碱样本集进行合理分级,采用粒子群算法直接优化投影寻踪的投影函数和投影方向,由最佳投影方向计算评价样本所属等级。在江苏滨海盐碱地暗管改碱效应评价中的应用表明,直接由样本数据驱动的投影寻踪等级评价模型对样本的综合效应等级评价合理,模型精度较高,平均相对误差为6.3%。

To evaluate the combined effect of coastal saline-alkali soil reclaimed by subsurface pipe drainage system, the index system was established from the three aspects of economic benefit, soil quality and ecological environment. By using the projection pursuit method which is based on the particle swarm optimization, the projection pursuit grade evaluation model of coastal saline-alkali soil reclaim was established, and its establishment process was introduced in detail. According to this model, the multi-dimensional grading indexes can be integrated into the one-dimensional projection value, and the samples can be reasonably graded according to the functional relation between the projection value and the evaluation grading experienced value. The particle swarm optimization arithmetic was applied to optimize the projection function and the direction of projection pursuit which will help to calculate the grade of the evaluating sample. Application of this model in evaluation of the effect of Jiangsu coastal saline-alkali soil reclaim shows that, this model derived by samples data is reasonable, the precision of the model is relatively high, and the mean relative error is 6. 3%.