首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于分类回归树算法的东南太平洋智利竹筴鱼渔场预报
引用本文:崔雪森,伍玉梅,张晶,周甦芳,樊伟.基于分类回归树算法的东南太平洋智利竹筴鱼渔场预报[J].中国海洋大学学报(自然科学版),2012(Z2):53-59.
作者姓名:崔雪森  伍玉梅  张晶  周甦芳  樊伟
作者单位:中国水产科学研究院东海水产研究所渔业资源遥感信息技术重点开放实验室
基金项目:国家高技术研究发展计划项目(2007AA092202)资助
摘    要:为了提高智利竹筴鱼渔场预报水平和满足渔业捕捞生产的需要,利用2002—2008年的东南太平洋公海海域捕捞的中国大型拖网渔船共计15艘的生产统计资料,以及海洋环境数据(包括海表温度、叶绿素a浓度、表温距平、叶绿素a浓度距平、海表温度梯度强度和海面高度异常等数据),基于CART的算法,构建了智利竹筴鱼渔场决策树预报模型。用含1 114条记录的数据集对模型进行训练,并采用ROC方法对该模型诊断中心渔场的准确性进行了分析。最后将该模型应用于2009年各月份的智利竹筴鱼中心渔场预报,并与实际渔场位置进行了对比,结果显示预报渔场与实际生产位置基本一致,表明利用CART决策树方法建立智利竹筴鱼渔场预报模型是可行的。

关 键 词:CART决策树  智利竹筴鱼  渔场预报  东南太平洋

Fishing Ground Forecasting of Chilean Jack Mackerel(Trachurus murphyi) in the Southeast Pacific Ocean Based on CART Decision Tree
CUI Xue-Sen,WU Yu-Mei,ZHANG Jing,ZHOU Su-Fang,FAN Wei.Fishing Ground Forecasting of Chilean Jack Mackerel(Trachurus murphyi) in the Southeast Pacific Ocean Based on CART Decision Tree[J].Periodical of Ocean University of China,2012(Z2):53-59.
Authors:CUI Xue-Sen  WU Yu-Mei  ZHANG Jing  ZHOU Su-Fang  FAN Wei
Institution:(The Key & Open Laboratory of Remote Sensing & Information Technology Application in Fisheries,East China Sea Fisheries Research Institute,Chinese Academy of Fishery Sciences,Shanghai 200090,China)
Abstract:To improve the forecasting accuracy of fishing ground and meet the needs of the practical fishery production,fishery statistics of 15 large-scale trawlers from China fishing in the open sea of Southeast Pacific and six oceanic environmental factors,including sea surface temperature(SST),Chlorophyll-a concentration(Chl-a),SST anomaly,Chl-a anomaly,SST gradient,sea surface height anomaly(SSHA) from 2002—2008 were applied to build the fishing ground prediction model for Trachurus murphyi based on CART decision tree.The model was trained by the dataset with 1 114 records.The accuracy of the model diagnosing fishing zone was analyzed by ROC method.The oceanic environmental dataset in 2009 were applied to test the effect of fishing forecast by the model in the practical application.Results showed that the predicted position of fishing grounds was consistent with that of practical fishing locations.It indicates that the application of CART decision tree method on Chilean jack mackerel fishing ground prediction was feasible.
Keywords:CART decision tree  Chilean jack mackerel  fishing zone prediction  Southeast Pacific
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号