采用迭代注水原理的居民地属性约简方法
|
摘要
基于案例推理的居民地自动选取方法研究中,居民地属性权重的赋值对推理结果的影响显著。为使案例推理中居民地属性权重分配更合理,引入迭代注水原理对居民地属性进行赋权值和约简。首先,对专家交互操作的面状居民地案例进行信息挖掘;然后采用注水原理对居民地属性权重优化分配,结合递归特征消除法对属性迭代计算;最后采用十折交叉验证法,训练出适应居民地案例推理的最佳赋有权重的属性子集。通过与专家打分法、主成分分析法、权重平均分配法、传统注水原理法4种属性约简方法作对比实验,结果表明本文方法能对居民地属性进行有效简约,并提高基于案例推理的居民地选取模型的正确率。
Among the research on the method of automatic selection of residents based on case-based reasoning(CBR),the influence of the attribute weights assignment of residents is significantly. In order to make the distribution of attribute weights more reasonable in CBR,the iteration water-filling theory is used for recursive feature elimination of weighting and reduction of attributes. Firstly,attribute mining is carried out on the case of residents selected by expert human-computer interaction,Then the water-filling theory is used to optimize the distribution of property weight of residents and combine with the recursive feature elimination method to calculate the attribute iteration. Finally,ten-fold cross validation method is used to train the attribute subset with optimal weights,which is adapted to the residents in CBR. By comparing the experiment results with those of four kinds of attribute reduction methods,which are namely expert evaluating method,analytic hierarchy process,mean assignment method,traditional water-filling theory,it shows that this method in this paper can effectively achieve the attribute reduction and improve the accuracy of the model based on CBR.
Among the research on the method of automatic selection of residents based on case-based reasoning(CBR),the influence of the attribute weights assignment of residents is significantly. In order to make the distribution of attribute weights more reasonable in CBR,the iteration water-filling theory is used for recursive feature elimination of weighting and reduction of attributes. Firstly,attribute mining is carried out on the case of residents selected by expert human-computer interaction,Then the water-filling theory is used to optimize the distribution of property weight of residents and combine with the recursive feature elimination method to calculate the attribute iteration. Finally,ten-fold cross validation method is used to train the attribute subset with optimal weights,which is adapted to the residents in CBR. By comparing the experiment results with those of four kinds of attribute reduction methods,which are namely expert evaluating method,analytic hierarchy process,mean assignment method,traditional water-filling theory,it shows that this method in this paper can effectively achieve the attribute reduction and improve the accuracy of the model based on CBR.
引文
[1]王家耀,姚杰.模糊综合评判在制图综合中的应用(以居民地选取为例)[J].解放军测绘学院学报,1985,12(2):45-52.WANG J Y,YAO J.An application of fuzzy comprehensive estimation in cartographic generalition[J].Journal of the PLA Institute of Science and Technology,1985,12(2):45-52.
[2]胡慧明,钱海忠,何海威,等.采用层次分析法的面状居民地自动选取[J].测绘学报,2016,45(6):740-746;755.HU H M,QIAN H Z,HE H W,et al.Auto-selection of areal habitation based on analytic hierarchy process[J].Acta Geodaetica et Cartographica Sinica,2016,45(6):740-746;755.
[3]何海威,钱海忠,刘闯,等.采用决策树算法进行居民地自动综合[J].测绘科学技术学报,2016,33(6):623-628.HE H W,QIAN H Z,LIU C,et al.Intelligent habitation generalization research by using decision tree algorithm[J].Journal of Geomatics Science and Technology,2016,33(6):623-628.
[4]SCHANK R C.Dynamic memory:A theory of reminding and learning in computers and people[M].Cambridge:Cambridge University Press,1983:29-57.
[5]PARK C S,HAN I.A case-based reasoning with the feature weights derived by analytic hierarchy process for bankruptcy prediction[J].Expert Systems with Applications,2002,23(3):255-264.
[6]AHN H,KIM K J,HAN I.Global optimization of feature weights and the number of neighbors that combine in a casebased reasoning system[J].Expert Systems,2006,23(5):290-301.
[7]PIAN J X,CHAI T Y,LI J J.Application of case-based reasoning and iterative learning to laminar cooling process control[J].Acta Automatica Sinica,2012,38(12):2032.
[8]赵辉,严爱军,王普.提高案例推理分类器的可靠性研究[J].自动化学报,2014,40(9):2029-2036.ZHAO H,YAN A J,WANG P.On improving reliability of case-based reasoning classifier[J].Acta Automatica Sinica,2014,40(9):2029-2036.
[9]严爱军,钱丽敏,王普.案例推理属性权重的分配模型比较研究[J].自动化学报,2014,40(9):1896-1902.YAN A J,QIAN L M,WANG P.A comparative study of attribute weights assignment for case-based reasoning[J].Acta Automatica Sinica,2014,40(9):1896-1902.
[10]胡慧明,钱海忠,何海威.采用主成分分析法的面状居民地自动选取[J].测绘与空间地理信息,2016,39(4):41-45;49.HU H M,QIAN H Z,HE H W.Auto-selection of areal habitation based on principal component analysis[J].Geomatics&Spatial Information Technology,2016,39(4):41-45;49.
[11]郑春燕,胡华科.居民地与道路之间拓扑关系一致性的模糊评价[J].测绘科学,2012,37(1):165-167.ZHENG C Y,HU H K.Fuzzy evaluation for the consistency of topological relations between residents and roads in multi-scale map representation[J].Science of Surveying and Mapping,2012,37(1):165-167.
[12]蔡永香,张成.居民地与其他线状要素间拓扑关系抽象的模糊推理[J].城市勘测,2006(5):10-14;17.CAI Y X,ZHANG C.Fuzzy reasoning of topological relations between residential area and line objects generalization[J].Urban Geotechnical Investigation&Surveying,2006(5):10-14;17.
[13]何宗宜.地图上确定居民地选取指标的依据研究[J].武汉测绘科技大学学报,1986,11(1):56-62.HE Z Y.A study of the basis for determing the quotas of settlements selected for mapmaking[J].Journal of Wuhan Technical University of Surveying and Mapping,1986,11(1):56-62.
[14]胡慧明.基于层次结构模型的居民地自动选取方法研究[D].郑州:信息工程大学,2016:12-45.HUI H M.Research on automatic habitation selection method based on hierarchical structure model[D].Zhengzhou:Information Engineering University,2016:12-45.
[15]蔡永香.基于Voronio图的居民地渐进式选取方法研究[J].长江大学学报(自科版),2007,4(1):66-68.CAI Y X.Study on the method of gradual selection of residents based on voronio graph[J].Journal of Yangtze University(Natural Science Edition),2007,4(1):66-68.
[16]宋鹰,何宗宜,粟卫民.基于Rough集的居民地属性知识约简与结构化选取[J].武汉大学学报(信息科学版),2005,30(4):329-332.SONG Y,HE Z Y,SU W M.Attributes reduction and structured selection in automatic cartographical generalization based on rough set[J].Geomatics and Information Science of Wuhan University,2005,30(4):329-332.
[2]胡慧明,钱海忠,何海威,等.采用层次分析法的面状居民地自动选取[J].测绘学报,2016,45(6):740-746;755.HU H M,QIAN H Z,HE H W,et al.Auto-selection of areal habitation based on analytic hierarchy process[J].Acta Geodaetica et Cartographica Sinica,2016,45(6):740-746;755.
[3]何海威,钱海忠,刘闯,等.采用决策树算法进行居民地自动综合[J].测绘科学技术学报,2016,33(6):623-628.HE H W,QIAN H Z,LIU C,et al.Intelligent habitation generalization research by using decision tree algorithm[J].Journal of Geomatics Science and Technology,2016,33(6):623-628.
[4]SCHANK R C.Dynamic memory:A theory of reminding and learning in computers and people[M].Cambridge:Cambridge University Press,1983:29-57.
[5]PARK C S,HAN I.A case-based reasoning with the feature weights derived by analytic hierarchy process for bankruptcy prediction[J].Expert Systems with Applications,2002,23(3):255-264.
[6]AHN H,KIM K J,HAN I.Global optimization of feature weights and the number of neighbors that combine in a casebased reasoning system[J].Expert Systems,2006,23(5):290-301.
[7]PIAN J X,CHAI T Y,LI J J.Application of case-based reasoning and iterative learning to laminar cooling process control[J].Acta Automatica Sinica,2012,38(12):2032.
[8]赵辉,严爱军,王普.提高案例推理分类器的可靠性研究[J].自动化学报,2014,40(9):2029-2036.ZHAO H,YAN A J,WANG P.On improving reliability of case-based reasoning classifier[J].Acta Automatica Sinica,2014,40(9):2029-2036.
[9]严爱军,钱丽敏,王普.案例推理属性权重的分配模型比较研究[J].自动化学报,2014,40(9):1896-1902.YAN A J,QIAN L M,WANG P.A comparative study of attribute weights assignment for case-based reasoning[J].Acta Automatica Sinica,2014,40(9):1896-1902.
[10]胡慧明,钱海忠,何海威.采用主成分分析法的面状居民地自动选取[J].测绘与空间地理信息,2016,39(4):41-45;49.HU H M,QIAN H Z,HE H W.Auto-selection of areal habitation based on principal component analysis[J].Geomatics&Spatial Information Technology,2016,39(4):41-45;49.
[11]郑春燕,胡华科.居民地与道路之间拓扑关系一致性的模糊评价[J].测绘科学,2012,37(1):165-167.ZHENG C Y,HU H K.Fuzzy evaluation for the consistency of topological relations between residents and roads in multi-scale map representation[J].Science of Surveying and Mapping,2012,37(1):165-167.
[12]蔡永香,张成.居民地与其他线状要素间拓扑关系抽象的模糊推理[J].城市勘测,2006(5):10-14;17.CAI Y X,ZHANG C.Fuzzy reasoning of topological relations between residential area and line objects generalization[J].Urban Geotechnical Investigation&Surveying,2006(5):10-14;17.
[13]何宗宜.地图上确定居民地选取指标的依据研究[J].武汉测绘科技大学学报,1986,11(1):56-62.HE Z Y.A study of the basis for determing the quotas of settlements selected for mapmaking[J].Journal of Wuhan Technical University of Surveying and Mapping,1986,11(1):56-62.
[14]胡慧明.基于层次结构模型的居民地自动选取方法研究[D].郑州:信息工程大学,2016:12-45.HUI H M.Research on automatic habitation selection method based on hierarchical structure model[D].Zhengzhou:Information Engineering University,2016:12-45.
[15]蔡永香.基于Voronio图的居民地渐进式选取方法研究[J].长江大学学报(自科版),2007,4(1):66-68.CAI Y X.Study on the method of gradual selection of residents based on voronio graph[J].Journal of Yangtze University(Natural Science Edition),2007,4(1):66-68.
[16]宋鹰,何宗宜,粟卫民.基于Rough集的居民地属性知识约简与结构化选取[J].武汉大学学报(信息科学版),2005,30(4):329-332.SONG Y,HE Z Y,SU W M.Attributes reduction and structured selection in automatic cartographical generalization based on rough set[J].Geomatics and Information Science of Wuhan University,2005,30(4):329-332.