摘要
针对生物量影响因子量化研究较少、方法单一及区域生物量评价不足且基于单个树种生物量模型进行评价时工作量过大的问题,以孟家岗林场的三类小班清查数据为基础,选取与生物量水平相关的11个因子,利用C5.0算法进行生物量决策树建模,并进一步利用Apriror算法进行生物量强影响因子的关联规则挖掘。结果表明:生物量决策树模型的分类预测精度为88.78%,生物量影响因子的量化结果分别为树高(0.348)、胸径(0.225)、林分类型(0.196)、龄级(0.162)、郁闭度(0.134)、坡度(0.096)、海拔(0.074)、坡向(0.065)、立地类型(0.052)和坡位(0.037);得到707条置信度在80%以上、支持度在10%以上的因子关联规则,揭示了生物量影响因子间的隐含关联关系。建立的生物量决策树模型能为快速的区域生物量预测和评价提供模型参考,建立的关联规则评估模型能够为以碳汇为目标的森林生产与经营提供客观评价指标。
Quantitative research on biomass impact factors was less, the used method was relatively single, the evaluation on regional biomass was inadequate and the biomass evaluating workload based on separated tree biomass models was overworked. Eleven factors related with biomass level were chosen based on the forest resources inventory data of Mengjiagang Forest Farm, the biomass decision tree model was built by using C5.0 algorithm, and further the correlation rules among factors which strongly influence the biomass level were developed by adopting Apriori algorithm. The results show that the classification accuracy of the biomass decision tree was 88.78%, the quantitative results of factors were as follows tree height(0.348), DBH(0.225), forest type(0.196), age class(0.162), canopy density(0.134), slope gradient(0.096), elevation(0.074), slope aspect(0.065), site type(0.052) and slope position(0.037); Seven hundred and seven correlation rules with confidence higher than 80% and support higher than 10% were obtained, which have revealed the hidden relations among the chosen factors. The conclusions come out that the biomass decision tree can provide a reference for quickly prediction and evaluation on regional biomass, the correlation rules can provide objective evaluation indexes for forest production and management under the goal of carbon sinks.
出处
《中南林业科技大学学报》
CAS
CSCD
北大核心
2015年第3期1-6,共6页
Journal of Central South University of Forestry & Technology
基金
国家"十二五"农村领域科技计划课题(2012AA102003-2)
中央高校基本科研业务费专项资金项目(2572014AB22)