论文部分内容阅读
针对黑龙江垦区各农场农机装备水平不平衡差异问题,采用支持向量机多类分类和主成分分析相结合的方法进行研究。将农机装备水平分为3个等级水平,从总量、速度和均量3方面选取10个评价指标;采用主成分分析法确定5个主要合成指标;建立多类分类支持向量机新模型,把新模型转化成一个互补问题,利用Lagrangian隐函数进一步转化成一个强凸的无约束优化问题,采取快速牛顿算法进行求解;利用实证调研数据,从发展差异度的角度对黑龙江垦区98个农场农机装备水平的差异进行多指标分析,在分类的准确度和训练速度方面都有很好的表现。
Aiming at the problem of unbalanced level of agricultural machinery and equipment in all the farms in Heilongjiang Reclamation, this paper uses the combination of support vector machine and multi-class classification and principal component analysis. The level of agricultural machinery equipment is divided into three levels, from the total amount, speed and average 3 selected 10 evaluation indicators; using principal component analysis to determine the five major synthetic indicators; to establish a new model of multi-class classification SVM, the The new model is transformed into a complementary problem, which is further transformed into a strong convex unconstrained optimization problem by using Lagrangian implicit function. A fast Newton algorithm is used to solve the problem. Using the data of empirical research, 98 farm farm machinery in Heilongjiang Reclamation Area Level differences for multi-index analysis, the classification accuracy and training speed has a good performance.