中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 地理科学与资源研究所
  3. 中国科学院地理科学与资源研究所
  4. 资源与环境信息系统国家重点实验室
  5. 外文论文
Comparing machine learning algorithms to predict vegetation fire detections in Pakistan

文献类型：期刊论文

作者Shahzad, Fahad7; Mehmood, Kaleem3,4; Hussain, Khadim2; Haidar, Ijlal3,4; Anees, Shoaib Ahmad5; Muhammad, Sultan4; Ali, Jamshid3; Adnan, Muhammad6; Wang, Zhichao7; Feng, Zhongke1,7
刊名FIRE ECOLOGY
出版日期2024-06-25
卷号20期号:1页码:20
关键词Machine learning Forest fire Crop fire Other vegetation fire Prediction models
ISSN号1933-9747
DOI10.1186/s42408-024-00289-5
英文摘要Vegetation fires have major impacts on the ecosystem and present a significant threat to human life. Vegetation fires consists of forest fires, cropland fires, and other vegetation fires in this study. Currently, there is a limited amount of research on the long-term prediction of vegetation fires in Pakistan. The exact effect of every factor on the frequency of vegetation fires remains unclear when using standard analysis. This research utilized the high proficiency of machine learning algorithms to combine data from several sources, including the MODIS Global Fire Atlas dataset, topographic, climatic conditions, and different vegetation types acquired between 2001 and 2022. We tested many algorithms and ultimately chose four models for formal data processing. Their selection was based on their performance metrics, such as accuracy, computational efficiency, and preliminary test results. The model's logistic regression, a random forest, a support vector machine, and an eXtreme Gradient Boosting were used to identify and select the nine key factors of forest and cropland fires and, in the case of other vegetation, seven key factors that cause a fire in Pakistan. The findings indicated that the vegetation fire prediction models achieved prediction accuracies ranging from 78.7 to 87.5% for forest fires, 70.4 to 84.0% for cropland fires, and 66.6 to 83.1% for other vegetation. Additionally, the area under the curve (AUC) values ranged from 83.6 to 93.4% in forest fires, 72.6 to 90.6% in cropland fires, and 74.2 to 90.7% in other vegetation. The random forest model had the highest accuracy rate of 87.5% in forest fires, 84.0% in cropland fires, and 83.1% in other vegetation and also the highest AUC value of 93.4% in forest fires, 90.6% in cropland fires, and 90.7% in other vegetation, proving to be the most optimal performance model. The models provided predictive insights into specific conditions and regional susceptibilities to fire occurrences, adding significant value beyond the initial MODIS detection data. The maps generated to analyze Pakistan's vegetation fire risk showed the geographical distribution of areas with high, moderate, and low vegetation fire risks, highlighting predictive risk assessments rather than historical fire detections. Los fuegos de vegetaci & oacute;n tienen grandes impactos en los ecosistemas y presentan una amenaza significativa para la vida humana. En este estudio, los fuegos de vegetaci & oacute;n comprenden fuegos forestales, en cultivos, y otros fuegos de vegetaci & oacute;n. Al presente, hay un limitado n & uacute;mero de investigaciones sobre la predicci & oacute;n a largo plazo de los fuegos de vegetaci & oacute;n en Pakist & aacute;n. El efecto exacto de cada factor en la frecuencia de los fuegos de vegetaci & oacute;n es poco claro cuando se usan an & aacute;lisis est & aacute;ndar. Esta investigaci & oacute;n utiliz & oacute; la alta eficiencia de los algoritmos del aprendizaje autom & aacute;tico (i. e. Machine Learning algorithms), para combinar datos de diversas fuentes, incluyendo datos del MODIS Global Fire Atlas, y datos topogr & aacute;ficos, de condiciones clim & aacute;ticas, y de diferentes tipos de vegetaci & oacute;n adquiridos entre 2001 y 2022. Probamos muchos algoritmos y finalmente elegimos cuatro modelos para procesar formalmente los datos. Su selecci & oacute;n fue basada en la performance de sus medidas, como la exactitud, eficiencia computacional, y los resultados preliminares de estas pruebas. El modelo de regresi & oacute;n log & iacute;stica, bosque al azar (random forest), un algoritmo de aprendizaje supervisado (support vector machine), y una t & eacute;cnica de potenciaci & oacute;n de gradiente extremo (extreme Gradient Boosting) fueron usados para identificar y elegir los nueve factores clave en fuegos forestales y en cultivos y, en caso de otro tipo de vegetaci & oacute;n, siete factores clave que causan incendios en Pakist & aacute;n. Los resultados indican que los modelos de predicci & oacute;n alcanzaron exactitudes que variaron entre 78,7 y el 87,5% para los fuegos forestales, el 70,4 al 84,0% en el caso de los fuegos en cultivos, y del 66,6 al 83,1% para otro tipo de vegetaci & oacute;n. Adicionalmente, el & aacute;rea de los valores bajo la curva (AUC) variaron del 83,6 al 93,4% para fuegos forestales, del 72,6 al 90,6% para los cultivos, y del 74,2 al 90,7% para otro tipo de vegetaci & oacute;n. El modelo Random Forest fue quien present & oacute; la mayor exactitud -87,5% en fuegos forestales, 84,0% en cultivos, y 83.1% en otro tipo de vegetaci & oacute;n-, y tambi & eacute;n el AUC m & aacute;s alto (93,4%) para fuegos forestales, (90,6%) en cultivos, y 90,7 en otro tipo de vegetaci & oacute;n, lo que prob & oacute; ser el modelo m & aacute;s & oacute;ptimo. Los modelos proveyeron de perspectivas predictivas en condiciones espec & iacute;ficas y susceptibilidades regionales a la ocurrencia de incendios, adicionando un valor significativo m & aacute;s all & aacute; de los datos iniciales de detecci & oacute;n por MODIS. Los mapas generados para analizar el riesgo de incendio de la vegetaci & oacute;n de Pakist & aacute;n mostraron & aacute;reas de distribuci & oacute;n geogr & aacute;fica con riesgo alto, moderado y bajo, se & ntilde;alando determinaciones predictivas m & aacute;s que detecciones hist & oacute;ricas de fuegos.
WOS关键词SPATIAL-PATTERNS ; BURNED AREA ; LOGISTIC-REGRESSION ; FOREST-FIRES ; RISK-ASSESSMENT ; NEURAL-NETWORK ; GIS ; SUSTAINABILITY ; PROVINCE ; DRIVERS
资助项目Beijing Forestry University[BLRC2023A03] ; Natural Science Foundation of Beijing[8232038] ; Natural Science Foundation of Beijing[8234065] ; Key Research and Development Projects of Ningxia Hui Autonomous Region[2023BEG02050]
WOS研究方向Environmental Sciences & Ecology ; Forestry
语种英语
WOS记录号WOS:001253213700001
出版者SPRINGER
资助机构Beijing Forestry University ; Natural Science Foundation of Beijing ; Key Research and Development Projects of Ningxia Hui Autonomous Region
源URL[http://ir.igsnrr.ac.cn/handle/311030/206219]  
专题资源与环境信息系统国家重点实验室_外文论文
通讯作者Wang, Zhichao; Feng, Zhongke
作者单位1.Hainan Univ, Coll Trop Crops, Key Lab Genet & Germplasm Innovat Trop Special For, Minist Educ, Haikou 570228, Peoples R China
2.Beijing Forestry Univ, State Forestry & Grassland Adm Key Lab Forest Reso, Beijing 100083, Peoples R China
3.Beijing Forestry Univ, Key Lab Silviculture & Conservat, Minist Educ, Conservat 100083, Peoples R China
4.Univ Swat, Inst Forest Sci, Main Campus Charbagh, Swat 19120, Pakistan
5.Univ Agr Dera Ismail Khan, Dept Forestry, Dera Ismail Khan 29050, Pakistan
6.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, State Key Lab Resources & Environm Informat Syst, Beijing 100101, Peoples R China
7.Beijing Forestry Univ, Precis Forestry Key Lab Beijing, Beijing 100083, Peoples R China
推荐引用方式
GB/T 7714		 Shahzad, Fahad,Mehmood, Kaleem,Hussain, Khadim,et al. Comparing machine learning algorithms to predict vegetation fire detections in Pakistan[J]. FIRE ECOLOGY,2024,20(1):20.
APA Shahzad, Fahad.,Mehmood, Kaleem.,Hussain, Khadim.,Haidar, Ijlal.,Anees, Shoaib Ahmad.,...&Feng, Zhongke.(2024).Comparing machine learning algorithms to predict vegetation fire detections in Pakistan.FIRE ECOLOGY,20(1),20.
MLA Shahzad, Fahad,et al."Comparing machine learning algorithms to predict vegetation fire detections in Pakistan".FIRE ECOLOGY 20.1(2024):20.

入库方式： OAI收割

来源：地理科学与资源研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。