基于Apriori算法的森林虫害预测方法1)
2017-09-03张子恺齐航王上蔡仕伟姚宇李丹
张子恺 齐航 王上 蔡仕伟 姚宇 李丹
(东北林业大学,哈尔滨,150040)
基于Apriori算法的森林虫害预测方法1)
张子恺 齐航 王上 蔡仕伟 姚宇 李丹
(东北林业大学,哈尔滨,150040)
针对现阶段对已出现森林虫害数据未能完成全面、及时地统计,以及难以准确预测森林虫害爆发的潜在外来诱因的问题,提出使用面向Web挖掘的主题网络爬虫搜集病虫害相关数据,并利用大数据挖掘频繁模式与关联规则的Apriori算法,挖掘结果得到满足最小支持度阈值的频繁2项集,并进一步从中筛选2种重要的特征子集,包括害虫与寄主之间的频繁模式,寄主与外来树种之间的频繁模式。解决了已出现的病虫害数据难以统计的难题;同时预测出针对某一地区害虫可能诱发森林虫害的外来树种。结果表明该方法能达到可靠、有效的森林虫害预测目的。
Apriori算法;频繁模式;特征子集;病虫害预测
//Journal of Northeast Forestry University,2017,45(8):93-96.
The experiment was conducted to solve the problem that the forest pest data failed to complete the comprehensive and timely statistics, and it is difficult to accurately predict the potential extrinsic incentive of forest pest outbreaks. It was proposed to use the web-based reptile for Web mining to collect pest and disease-related data, and use large data Apriori algorithm of mining frequent patterns and association rules to extract the frequent 2 itemsets in satisfying the minimum support thresholds, and two important feature subsets were selected including the frequent patterns between pests and hosts, and the host and alien species between the frequent patterns. The experiment solved the pests and diseases in statistics, and predicted the alien species from the induced forest pests in a certain area.
森林虫害是我国森林灾害中常见的生物灾害,我国林木种类丰富,所以森林虫害种类很多,实现统计与防治有较大的难度。同时又普遍存在区域性的树种引进种植现象,使得潜在虫害爆发的可能性增大。Nair K S S[1]对引发外来树种虫害的虫源因素和导致虫害爆发的原因进行了研究,结果显示:当在本地林业生态系统中引进外来树种,特别是当害虫遭遇与其原寄主亲缘关系较近的树种,例如同属内的不同树种时,虫害的爆发风险极大。所以,提取害虫、寄主和同属树种间的相关性十分重要。Apriori算法[2]是数据挖掘领域挖掘频繁模式及发掘关联规则的重要应用算法。……
