基于机器学习算法的湖南沅陵森林火灾风险预测

打开文本图片集

中图分类号：S763.2 文献标志码：A 文章编号：1673-923X（2025）10-0059-10

Machine learning algorithm-based forest fire risk prediction in Yuanling， Hunan Province

Abstract：【Ojetie】Indertouratelyssstheskleveloffrestfireshelprestpatroloptizersouceltad improvefireprevetecyYuangountyakeasthseahectasedoeataofain，fueleloical andhumaactivityactors，macinelingalgoriasuedtouildafrestfccueepredictiomodel，ichsan reference significanceforforestfireprevention.【Method】Withcomprehensiveconsiderationofterrain，fuel，meteorologynd humanactivities，1ivingfactorsereetractedinesdyreaincdingelevationslope，slopediecton，omalidetatio index，vegeaioiaioieddroodsoidetle fromwatersystem，andthedrivingfactorsofforestfrewereevaluated.Thehistoricalfirepointdataofthestudyareaisobtained basedonthe MODIS fireproduct.Theforestfreoccurrnceprediction modelwasconstructed byamachinelearningalgorithm，and thepredictionaccuracyoftemodel wascomprehensivelyevaluatedbyusing theconfusionmatrix evaluationindexandtheROC curve.Theforestfireocurrencepredictionmodelwasconstructedbyamachneleaingalgorithm，andthepredictionaccuracyof themodelwascomprehensivelyevaluatedbyusingtheconfusionmatrixevaluationindexandtheROCcurve.【Result】Thedistance fromtheroadandthedistancefromtheresidentialareaaretwodrivingfactorswiththelargestweight，andotherdrivingfactorsalso affecttheocuenceofforestfresTheOCcurvesoftreemodelssowedtateadomforestmodeladoodaccuracyiththe accuracy reaching 78.15% andthe area under the curve 0.85，while the logistic regresson prediction model had the accuracy reaching 74.81% and the area under the curve 0.81.The accuracy of the SVM prediction model is 70.74% ，and the area under the curve is 0.79. 【Conclusion】Therandomforestmodelshowsbeterpredictionabilitythanthelogisticregressionmodelandsupportvectormacine model.Theareaswith highand extremelyhigh riskof forest fireaccounted for 26.62% inthestudyarea.The forest firerisk level map is helpfulfortherelevantdepartmentstotakerelevantpreventivemeasuresandeffectivelyprotectthesafetyofforestresources.

eywords：machine learning;random forest;supportvector machine;firerisk;predictionmodel;driving factol

森林具备多重生态服务功能，在气候调节、水源涵养、风沙防护、土壤保持以及固碳释氧等方面发挥着无可替代的作用[1]。（剩余13626字）