東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Novel classification techniques for ...

Liu, Chuangmin.

FindBook

Google Book

Amazon

博客來

Novel classification techniques for tree species and ecological habitats in mixed-species forests.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Novel classification techniques for tree species and ecological habitats in mixed-species forests./
作者:	Liu, Chuangmin.
面頁冊數:	133 p.
附註:	Major Professor: Lianjun Zhang.
Contained By:	Dissertation Abstracts International64-03B.
標題:	Agriculture, Forestry and Wildlife. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3084754

Novel classification techniques for tree species and ecological habitats in mixed-species forests.
Liu, Chuangmin.

Novel classification techniques for tree species and ecological habitats in mixed-species forests. - 133 p.

Major Professor: Lianjun Zhang.

Thesis (Ph.D.)--State University of New York College of Environmental Science and Forestry, 2003.

Mixed-species forest stands consist of multiple species with different growth rates and shade tolerances, thus are characterized by different ages and distributions of diameter at breast height (dbh). Over a long time and large area, mixed-species stands form ecological habitats. Each has different productivity and regeneration, and needs different management strategies. Therefore, accurate classification is essential and critical within stand and at a landscape level. This study applied novel classification methods to describe the dbh distributions of mixed-species stands and to classify Forest Inventory and Analysis (FIA) plots into ecological habitat types in the Northeast, USA. The results showed: (1) Within a single mixed species stand, the finite mixture model produced much smaller root mean square error and bias, and fitted the entire distribution of the plots with extreme peaks, bimodality or heavy-tails better, as compared with traditional methods in which a single Weibull function was fit to either the whole plot or each species component separately. (2) The artificial neural network (ANN) models outperformed the traditional statistical methods such as linear discriminant analysis and minimum-distance classification method. The classification accuracy of the ANN models was 90% or higher for overall classification, and exceeded 92% in five of the six habitat categories. (3) A membership function was developed to further improve the classification of ambiguous plots with mixed overstory and understory species compositions. The classification accuracies of fuzzy c-means methods and multilayer perceptron method using the membership function were 98% and 97%, respectively, for overall classification. The novel classification techniques were shown to be attractive alternatives to traditional quantitative techniques. These more accurate methods for characterizing the diameter distributions within mixed-species stands and classifying the mixed-species stands into ecological habitats are very useful in the development of decision-support systems for forest resources management.Subjects--Topical Terms:

783690
Agriculture, Forestry and Wildlife.

Novel classification techniques for tree species and ecological habitats in mixed-species forests.
LDR:03080nam 2200277 a 45 001 937630
005 20110511
008 110511s2003 eng d
035 $a (UnM)AAI3084754
035 $a AAI3084754
040 $a UnM $c UnM
100 1 $a Liu, Chuangmin. $3 1261493
245 1 0 $a Novel classification techniques for tree species and ecological habitats in mixed-species forests.
300 $a 133 p.
500 $a Major Professor: Lianjun Zhang.
500 $a Source: Dissertation Abstracts International, Volume: 64-03, Section: B, page: 1020.
502 $a Thesis (Ph.D.)--State University of New York College of Environmental Science and Forestry, 2003.
520 $a Mixed-species forest stands consist of multiple species with different growth rates and shade tolerances, thus are characterized by different ages and distributions of diameter at breast height (dbh). Over a long time and large area, mixed-species stands form ecological habitats. Each has different productivity and regeneration, and needs different management strategies. Therefore, accurate classification is essential and critical within stand and at a landscape level. This study applied novel classification methods to describe the dbh distributions of mixed-species stands and to classify Forest Inventory and Analysis (FIA) plots into ecological habitat types in the Northeast, USA. The results showed: (1) Within a single mixed species stand, the finite mixture model produced much smaller root mean square error and bias, and fitted the entire distribution of the plots with extreme peaks, bimodality or heavy-tails better, as compared with traditional methods in which a single Weibull function was fit to either the whole plot or each species component separately. (2) The artificial neural network (ANN) models outperformed the traditional statistical methods such as linear discriminant analysis and minimum-distance classification method. The classification accuracy of the ANN models was 90% or higher for overall classification, and exceeded 92% in five of the six habitat categories. (3) A membership function was developed to further improve the classification of ambiguous plots with mixed overstory and understory species compositions. The classification accuracies of fuzzy c-means methods and multilayer perceptron method using the membership function were 98% and 97%, respectively, for overall classification. The novel classification techniques were shown to be attractive alternatives to traditional quantitative techniques. These more accurate methods for characterizing the diameter distributions within mixed-species stands and classifying the mixed-species stands into ecological habitats are very useful in the development of decision-support systems for forest resources management.
590 $a School code: 0213.
650 4 $a Agriculture, Forestry and Wildlife. $3 783690
650 4 $a Computer Science. $3 626642
650 4 $a Statistics. $3 517247
690 $a 0463
690 $a 0478
690 $a 0984
710 2 0 $a State University of New York College of Environmental Science and Forestry. $3 1017822
773 0 $t Dissertation Abstracts International $g 64-03B.
790 $a 0213
790 1 0 $a Zhang, Lianjun, $e advisor
791 $a Ph.D.
792 $a 2003
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3084754