Data Set Information:

仅从制图变量预测森林覆盖类型（无遥感数据）。根据美国林业局（USFS）区域2资源信息系统（RIS）数据确定给定观测（30 x 30米单元）的实际森林覆盖类型。自变量来自最初从美国地质调查局（USGS）和USFS数据获得的数据。数据为原始形式（未缩放），包含定性自变量（荒野区域和土壤类型）的二进制（0或1）数据列。

该研究区域包括位于科罗拉多州北部罗斯福国家森林的四个荒野区域。这些地区代表的是人为干扰最小的森林，因此现有的森林覆盖类型更多地是生态过程的结果，而不是森林管理做法。

这四个荒野地区的一些背景信息：在这四个荒野地区中，Neota（地区2）的平均海拔值可能最高。Rawah（1区）和Comanche Peak（3区）的平均海拔值较低，而Cache la Poudre（4区）的平均海拔值最低。

至于这些地区的主要主要树种，Neota的主要树种为云杉/冷杉（类型1），而Rawah和Comanche Peak的主要树种可能为罗奇波尔松（类型2），其次为云杉/冷杉和白杨（类型5）。卡什拉普德雷往往有黄松（类型3）、道格拉斯冷杉（类型6）和棉白杨/柳树（类型4）。

Rawah和Comanche山顶区域往往比Neota或Cache la Poudre更为典型，因为它们的树种分类和预测变量值范围（海拔等）。Cache la Poudre可能比其他区域更为独特，由于其相对较低的海拔范围和物种组成。

Attribute Information:

给出了属性名称、属性类型、度量单位和简要说明。森林覆盖类型是分类问题。此列表的顺序对应于数据库行中数字的顺序。

Name / Data Type / Measurement / Description

Elevation / quantitative /meters / Elevation in meters
Aspect / quantitative / azimuth / Aspect in degrees azimuth
Slope / quantitative / degrees / Slope in degrees
Horizontal_Distance_To_Hydrology / quantitative / meters / Horz Dist to nearest surface water features
Vertical_Distance_To_Hydrology / quantitative / meters / Vert Dist to nearest surface water features
Horizontal_Distance_To_Roadways / quantitative / meters / Horz Dist to nearest roadway
Hillshade_9am / quantitative / 0 to 255 index / Hillshade index at 9am, summer solstice
Hillshade_Noon / quantitative / 0 to 255 index / Hillshade index at noon, summer soltice
Hillshade_3pm / quantitative / 0 to 255 index / Hillshade index at 3pm, summer solstice
Horizontal_Distance_To_Fire_Points / quantitative / meters / Horz Dist to nearest wildfire ignition points
Wilderness_Area (4 binary columns) / qualitative / 0 (absence) or 1 (presence) / Wilderness area designation
Soil_Type (40 binary columns) / qualitative / 0 (absence) or 1 (presence) / Soil Type designation
Cover_Type (7 types) / integer / 1 to 7 / Forest Cover Type designation

Relevant Papers:

Blackard, Jock A. and Denis J. Dean.  2000.  "Comparative Accuracies of Artificial Neural Networks and Discriminant Analysis in Predicting Forest Cover Types from Cartographic Variables."  Computers and Electronics in Agriculture  24(3):131-151.
[Web link]

Blackard, Jock A. and Denis J. Dean.  1998.  "Comparative Accuracies of Neural Networks and Discriminant Analysis in Predicting Forest Cover Types from Cartographic Variables."  Second Southern Forestry GIS Conference. University of Georgia.  Athens, GA.  Pages 189-199.

Blackard, Jock A.  1998.  "Comparison of Neural Networks and Discriminant Analysis in Predicting Forest Cover Types." Ph.D. dissertation.  Department of Forest Sciences.  Colorado State University.  Fort Collins, Colorado.  165 pages.

Papers That Cite This Data Set¹:

Joao Gama and Ricardo Rocha and Pedro Medas. Accurate decision trees for mining high-speed data streams. KDD. 2003.  [View Context].

Nikunj C. Oza and Stuart J. Russell. Experimental comparison

Original Owners of Database:

Remote Sensing and GIS Program
Department of Forest Sciences
College of Natural Resources
Colorado State University
Fort Collins, CO  80523
(contact Jock A. Blackard, jblackard '@' fs.fed.us or Dr. Denis J. Dean, denis.dean '@' utdallas.edu)

Donors of database:

1. Jock A. Blackard (jblackard '@' fs.fed.us)
GIS Coordinator
USFS - Forest Inventory & Analysis
Rocky Mountain Research Station
507 25th Street
Ogden, UT 84401

2. Dr. Denis J. Dean (denis.dean '@' utdallas.edu)
Professor
Program in Geography and Geospatial Sciences
School of Economic, Political and Policy Sciences
800 West Campbell Rd
Richardson, TX  75080-3021

3. Dr. Charles W. Anderson (anderson '@' cs.colostate.edu)
Associate Professor
Department of Computer Science
Colorado State University
Fort Collins, CO  80523  USA