支持向量机参数优化的土地集约利用评价与分析

Evaluation and Analysis of Land Intensive Utilization Based on Parameters Optimization of SVM

对国内外土地集约利用评价的相关文献研究,在支持向量机、蚁群算法基础上,提出相关系数、蚁群算法与支持向量机相结合评价方法,对指标进行相关分析,确定指标集,运用蚁群算法,优化支持向量机参数,得出较好的惩罚因子C,核函数σ和不敏感系数ε,再对支持向量机训练,该方法提高了训练准确度,对土地集约利用进行c ACO-SVM评价,并与ACO-SVM、GA-SVM的土地集约利用评价进行比较,评价与仿真结果表明,c ACO-SVM的土地集约利用评价优于ACO-SVM、GA-SVM两种方法,c ACO-SVM的土地集约利用评价效果比较理想。

Based on relevant literature research of evaluation on intensive land-use both at home and abroad, the theory of Support Vector Machine （SVM） and Ant Colony Algorithm （ACO） was discussed. A new method of Correlation Coefficient, the Ant Colony Algorithm and Support Vector Machine （cACO-SVM） was proposed, which analyzed the relevant indicators to determine index set, using ACO, optimization of SVM parameters to draw a good penalty factor C and kernel function sigma and epsilon insensitive coefficient and training SVM, the method improved the training accuracy. Optimization of the land intensive utilization evaluation based on cACO-SVM was put forward, comparing with the ACO-SVM and GA - SVM intensive land use evaluation. Evaluation and simulation results show that analysis of cACO-SVM intensive land use evaluation is better than that of the ACO - SVM and GA - two methods of SVM. Intensive land use evaluation effect of cACO - SVM is more ideal.