基于粗糙集C4.5算法的应用  

Application of C4.5 algorithm based on rough set

作　　者：贾晨光[1] 陈桂芬[1] 

机构地区：[1]吉林农业大学信息技术学院,长春市130118

出　　处：《中国农机化学报》2016年第7期149-153,共5页Journal of Chinese Agricultural Mechanization

基　　金：国家星火计划(2008GA661003);吉林省农委项目(201105068)

摘　　要：权重确定是管理决策问题研究中最基本的问题,最常见的属性权重通常依赖于专家的经验知识而确立,存在一定的不确定性。运用粗糙集理论的相对正域概念,通过调整客观权重和主观权重所占的比例进行权重分配分析,并结合C4.5算法对农安县土壤养分数据进行分类处理。研究结果表明:(1)通过可视化图形对比形式得出引入粗糙集权重的C4.5算法分类效果较C4.5算法更明显;(2)有分类结果可知:经过连续多年的精准施肥,农安县耕地中高肥力土壤逐渐上升,土壤肥力呈均匀分布。这一结果与农安县当地肥力情况较符,说明基于粗糙集的C4.5算法是一种有效的土壤肥力评价方法。In research of management decision,weight determination is the most basic problem.And the most common attribute weights often rely on knowledge of the experts.So,there is always some uncertainty.It is used relative positive region of rough set theory,by adjusting the proportion of objective weights and subjective weight to analyze weight distribution.Simultaneously,it is combined with C4.5algorithm to classify soil nutrient data process,which came from Nongan town.The results showed that the classification effect of rough set weight C4.5algorithm is more obvious than C4.5algorithm inferred by graphical comparison.And,it be seen from the classification results,after several years of precise fertilization,gradual increase of middle and high fertility and distributed evenly of soil fertility.This result is consistent with local fertility situation of Nongan,which showed that rough set C4.5algorithm is an effective method to evaluate soil fertility.

关 键 词：属性权重分配分析 粗糙集C4.5算法 土壤肥力分级 精准施肥 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]