简介:Thisstudywasonsuperiorityofthenon-negativematrixfactorization(NMF)algorithmforapplicationofinformationextractedwithaerialimages.First,NMFwasusedforaerialimageinformationextraction,andthenthisdatawascomparedwithaprincipalcomponentanalysis(PCA)inwhichr(thenumberofrowsorcolumnsofbasicmatrix)andEignum(thenumberofeigenvalues)weregivendifferentvalues.ExperimentalresultsshowedthattheruntimeofNMFwithr=20or50waslessthanthatofPCAwithanEignum=20or50.Also,therecognitionrateofNMFwithr=50washigherthanthatofanEignum=50.Theexperimentshowedthatnonnegativematrixfactorizationhadadvantagesofashorttimeperiodwithahighrecognitionrate.
简介:APoissonregressionmodelandanegativebinomialregressionmodel(NBmodel)areoftenusedinareassuchasmedicineandeconomy,butrarelyinthedomesticforestrysector,especiallyintheforestfireforecasting.BasedonthedataofforestfireoccurrencesinDaxing’anlingregionin1980-2005,thispaperprofoundlyanalyzestheapplicationconditionsandtestmethodsofthetwomodels.TheAICmethodwasusedtocheckthefittinglevelofthemodelsandthecapabilityofthemodelsforforecastingforestfireswasdiscussed.ThisstudyprovidednecessarytheoreticalbasisanddatasupportfortheapplicationofthetwomodelsinthefieldofforestryinChina.