Insurance claim level estimation and influencing factor analysis for electric vehicles: A case study in Changsha city

{"OBJECTIVES:":[0],"Claim":[1,204,206,209],"amount":[2,47,147],"levels":[3,48],"reflect":[4],"severity":[5],"risk":[6,9,17,261],"and":[7,69,92,131,144,158,181,208,226,257],"facilitates":[8],"estimation":[10],"in":[11,19,49,116,128,263],"insurance":[12,22],"practice.":[13],"This":[14,52],"study":[15,53],"supports":[16],"management":[18],"electric-vehicle":[20],"traffic":[21,50],"by":[23],"establishing":[24],"an":[25,172],"effective":[26],"classification":[27,102,154],"framework":[28],"for":[29],"claim":[30,46,109,146,245],"levels.":[31,148],"METHODS:":[32],"The":[33,108,119,135,191],"Synthetic":[34],"Minority":[35],"Over-Sampling":[36],"Technique":[37],"(SMOTE)":[38],"is":[39,86,165],"applied":[40,87],"to":[41,88,141,235],"balance":[42],"the":[43,76,101,123,151,161,168,186,199,232,238,243,251],"distribution":[44,130],"of":[45,103,153,174,179,189,219,253],"insurance.":[51,266],"applies":[54],"six":[55],"distinct":[56],"models-Decision":[57],"Tree,":[58],"Random":[59,136,162],"Forest,":[60],"CatBoost,":[61],"Support":[62],"Vector":[63],"Machine":[64,73],"(SVM),":[65],"Multilayer":[66],"Perceptron":[67],"(MLP),":[68],"Light":[70],"Gradient":[71],"Boosting":[72],"(LGBM)-to":[74],"classify":[75],"\\"Claim":[77,104],"Amount":[78,105],"Level.\\"":[79,106],"Finally,":[80],"SHapley":[81],"Additive":[82],"exPlanations":[83],"(SHAP)":[84],"method":[85],"provide":[89,258],"both":[90],"local":[91],"global":[93],"explanations,":[94],"thereby":[95],"elucidating":[96],"how":[97],"specific":[98],"features":[99,143],"influence":[100],"RESULTS:":[107],"data":[110,125],"were":[111],"collected":[112],"from":[113],"electric":[114,264],"vehicles":[115],"Changsha,":[117],"China.":[118],"results":[120],"show":[121],"that":[122],"enhanced":[124],"maintains":[126],"consistency":[127],"feature":[129],"improves":[132],"model":[133,138,164],"stability.":[134],"Forest":[137,163],"was":[139],"employed":[140],"select":[142],"estimate":[145],"Based":[149],"on":[150,198],"evaluation":[152],"performance,":[155],"class":[156],"separability,":[157],"business":[159],"interpretability,":[160],"identified":[166],"as":[167],"optimal":[169],"classifier,":[170],"achieving":[171],"accuracy":[173],"0.831,":[175],"a":[176,182],"macro-average":[177,183],"F1-score":[178],"0.826,":[180],"area":[184],"under":[185],"curve":[187],"(AUC)":[188],"0.950.":[190],"top":[192],"three":[193],"most":[194],"important":[195],"features,":[196],"based":[197],"SHAP":[200],"analysis,":[201],"are":[202],"Injury":[203],"Flag,":[205],"Type,":[207],"Count":[210],"During":[211],"Policy.":[212],"CONCLUSIONS:":[213],"SHAP-based":[214],"analysis":[215],"reveals":[216],"heterogeneous":[217],"effects":[218],"key":[220],"predictors.":[221],"Driver":[222],"behavioral":[223],"faults":[224],"(negligence":[225],"insufficient":[227],"safety":[228],"distance)":[229],"interact":[230],"with":[231],"injury":[233],"status":[234],"further":[236],"elevate":[237],"model's":[239],"predicted":[240],"probability":[241],"within":[242],"high":[244],"level":[246],"context.":[247],"These":[248],"findings":[249],"enhance":[250],"interpretability":[252],"machine":[254],"learning":[255],"models":[256],"insights":[259],"into":[260],"differentiation":[262],"vehicle":[265]}