tb-vulnerability-hotspot-predictor / ablation_results.json
t6harsh's picture
Upload ablation_results.json with huggingface_hub
d3ebb42 verified
Raw
History Blame Contribute Delete
18.3 kB
{
"feature_groups": {
"all_features": {
"accuracy": 0.934156378600823,
"accuracy_std": 0.0033600682342704623,
"f1_macro": 0.9025098817579865,
"f1_macro_std": 0.0032711375359027986,
"f1_weighted": 0.9351707425320005,
"f1_weighted_std": 0.00252075341300129,
"precision": 0.8150491476919249,
"precision_std": 0.04058853585786643,
"recall": 0.8866666666666667,
"recall_std": 0.04988876515698589,
"auc_roc": 0.9712953367875649,
"auc_roc_std": 0.0030752361124773545,
"ece": 0.05625765306190419,
"ece_std": 0.008588383909395743,
"brier_score": 0.05581313623709779,
"brier_score_std": 0.006274220148221518
},
"drop_demographic": {
"accuracy": 0.9231824417009603,
"accuracy_std": 0.007759745198206247,
"f1_macro": 0.8859667578371555,
"f1_macro_std": 0.013763613216983155,
"f1_weighted": 0.9242796237468868,
"f1_weighted_std": 0.008295054350748773,
"precision": 0.7880747126436782,
"precision_std": 0.022281698878262694,
"recall": 0.86,
"recall_std": 0.058878405775518984,
"auc_roc": 0.9731606217616582,
"auc_roc_std": 0.0017768319377702078,
"ece": 0.052618936281241736,
"ece_std": 0.0026563482885394213,
"brier_score": 0.0571605721833377,
"brier_score_std": 0.003969568180693764
},
"drop_nutritional": {
"accuracy": 0.9300411522633745,
"accuracy_std": 0.00888990493608756,
"f1_macro": 0.89688575862919,
"f1_macro_std": 0.010190860509204883,
"f1_weighted": 0.9312651519062315,
"f1_weighted_std": 0.007753314441575472,
"precision": 0.8051479579359256,
"precision_std": 0.0506662122098875,
"recall": 0.88,
"recall_std": 0.043204937989385774,
"auc_roc": 0.9747841105354058,
"auc_roc_std": 0.005028725120655431,
"ece": 0.05112197499182398,
"ece_std": 0.010045868219477016,
"brier_score": 0.055542791170826326,
"brier_score_std": 0.007900172699135828
},
"drop_socioeconomic": {
"accuracy": 0.9272976680384089,
"accuracy_std": 0.003879872599103124,
"f1_macro": 0.8913635596802715,
"f1_macro_std": 0.005740032291453515,
"f1_weighted": 0.928077016260845,
"f1_weighted_std": 0.0032616166857185867,
"precision": 0.8086979371961635,
"precision_std": 0.05349863273069293,
"recall": 0.86,
"recall_std": 0.07118052168020875,
"auc_roc": 0.9725043177892919,
"auc_roc_std": 0.0022865864129098827,
"ece": 0.058421829055497386,
"ece_std": 0.007957945257837242,
"brier_score": 0.05898473729251519,
"brier_score_std": 0.005418553338063918
},
"drop_environmental": {
"accuracy": 0.9272976680384089,
"accuracy_std": 0.01026517801584072,
"f1_macro": 0.892875007730843,
"f1_macro_std": 0.014588942717511452,
"f1_weighted": 0.9285557078393758,
"f1_weighted_std": 0.009728794325890293,
"precision": 0.7970856102003644,
"precision_std": 0.0531401839574474,
"recall": 0.88,
"recall_std": 0.0748331477354788,
"auc_roc": 0.9717098445595855,
"auc_roc_std": 0.004798274363850904,
"ece": 0.05447866720518955,
"ece_std": 0.009101877611029157,
"brier_score": 0.057996770037197554,
"brier_score_std": 0.00644363193679079
},
"drop_healthcare": {
"accuracy": 0.9300411522633745,
"accuracy_std": 0.0033600682342704623,
"f1_macro": 0.8966926208735324,
"f1_macro_std": 0.002801841175856656,
"f1_weighted": 0.9312113571659425,
"f1_weighted_std": 0.0024363943761779557,
"precision": 0.8032299175271121,
"precision_std": 0.03605109624893552,
"recall": 0.88,
"recall_std": 0.043204937989385774,
"auc_roc": 0.969706390328152,
"auc_roc_std": 0.0032817016584493425,
"ece": 0.053590356701218166,
"ece_std": 0.007088597700711033,
"brier_score": 0.05779605248571739,
"brier_score_std": 0.004837322186373437
},
"only_demographic": {
"accuracy": 0.7764060356652949,
"accuracy_std": 0.01180017183407767,
"f1_macro": 0.6323504875342443,
"f1_macro_std": 0.019841171533820322,
"f1_weighted": 0.7677400635488437,
"f1_weighted_std": 0.011877228507686981,
"precision": 0.4476694085389738,
"precision_std": 0.035858526683525724,
"recall": 0.3666666666666667,
"recall_std": 0.0410960933531265,
"auc_roc": 0.7070466321243524,
"auc_roc_std": 0.029369174916989992,
"ece": 0.13712007752964459,
"ece_std": 0.02363188559702372,
"brier_score": 0.17116552768530777,
"brier_score_std": 0.01236118363590424
},
"only_nutritional": {
"accuracy": 0.7256515775034295,
"accuracy_std": 0.010801108194803566,
"f1_macro": 0.5233204157524848,
"f1_macro_std": 0.013679053837226465,
"f1_weighted": 0.7060670891377628,
"f1_weighted_std": 0.009611705058173494,
"precision": 0.2605225546402017,
"precision_std": 0.030232197564110358,
"recall": 0.18000000000000002,
"recall_std": 0.016329931618554526,
"auc_roc": 0.5682556131260795,
"auc_roc_std": 0.037280007784803004,
"ece": 0.16882323983456385,
"ece_std": 0.015388775332815376,
"brier_score": 0.2034708341421735,
"brier_score_std": 0.007328561457780753
},
"only_socioeconomic": {
"accuracy": 0.7133058984910837,
"accuracy_std": 0.010265178015840709,
"f1_macro": 0.5216959726278539,
"f1_macro_std": 0.04387223461155405,
"f1_weighted": 0.6994907842935604,
"f1_weighted_std": 0.017493171442481675,
"precision": 0.23970773494583017,
"precision_std": 0.07290131771540098,
"recall": 0.20666666666666667,
"recall_std": 0.09843215373488932,
"auc_roc": 0.5621761658031087,
"auc_roc_std": 0.043979189964741386,
"ece": 0.16175159224951338,
"ece_std": 0.019751401070095608,
"brier_score": 0.20407348577598647,
"brier_score_std": 0.00760765222174475
},
"only_environmental": {
"accuracy": 0.7914951989026063,
"accuracy_std": 0.013579554096860985,
"f1_macro": 0.6773251437928417,
"f1_macro_std": 0.026367182289513465,
"f1_weighted": 0.7902616650585124,
"f1_weighted_std": 0.015415521788714043,
"precision": 0.49154391707583195,
"precision_std": 0.03317576823897603,
"recall": 0.48,
"recall_std": 0.05656854249492383,
"auc_roc": 0.7855958549222798,
"auc_roc_std": 0.019820125923107293,
"ece": 0.1257641581498692,
"ece_std": 0.01335097764422107,
"brier_score": 0.15604989863572077,
"brier_score_std": 0.010250038894142643
},
"only_healthcare": {
"accuracy": 0.7160493827160495,
"accuracy_std": 0.010080204702811431,
"f1_macro": 0.5357954081367066,
"f1_macro_std": 0.041127627746483555,
"f1_weighted": 0.705852665512889,
"f1_weighted_std": 0.018763680460714468,
"precision": 0.26566951566951563,
"precision_std": 0.06228020580825975,
"recall": 0.2333333333333333,
"recall_std": 0.09285592184789412,
"auc_roc": 0.5577892918825561,
"auc_roc_std": 0.015592861350262787,
"ece": 0.18157288400552682,
"ece_std": 0.006434572376156912,
"brier_score": 0.2062071678838663,
"brier_score_std": 0.0007350408851747545
}
},
"spatial_encoding": {
"no_spatial": {
"accuracy": 0.934156378600823,
"accuracy_std": 0.0033600682342704623,
"f1_macro": 0.9025098817579865,
"f1_macro_std": 0.0032711375359027986,
"f1_weighted": 0.9351707425320005,
"f1_weighted_std": 0.00252075341300129,
"precision": 0.8150491476919249,
"precision_std": 0.04058853585786643,
"recall": 0.8866666666666667,
"recall_std": 0.04988876515698589,
"auc_roc": 0.9712953367875649,
"auc_roc_std": 0.0030752361124773545,
"ece": 0.05625765306190419,
"ece_std": 0.008588383909395743,
"brier_score": 0.05581313623709779,
"brier_score_std": 0.006274220148221518
},
"spatial_lag_only": {
"accuracy": 0.934156378600823,
"accuracy_std": 0.0033600682342704623,
"f1_macro": 0.9015887525644697,
"f1_macro_std": 0.002713465695110215,
"f1_weighted": 0.9348745324986095,
"f1_weighted_std": 0.002304215818872965,
"precision": 0.8238220071250923,
"precision_std": 0.04603309044329161,
"recall": 0.8733333333333334,
"recall_std": 0.0524933858267454,
"auc_roc": 0.9810017271157169,
"auc_roc_std": 0.0020470511074306354,
"ece": 0.04948789444410349,
"ece_std": 0.005827174629755799,
"brier_score": 0.05093533682449882,
"brier_score_std": 0.004620922764351714
},
"fourier_only": {
"accuracy": 0.9368998628257886,
"accuracy_std": 0.01026517801584072,
"f1_macro": 0.906585065329408,
"f1_macro_std": 0.017412547900051915,
"f1_weighted": 0.9378711419036461,
"f1_weighted_std": 0.010758266021661871,
"precision": 0.8156041986687148,
"precision_std": 0.02938241171529233,
"recall": 0.9,
"recall_std": 0.07118052168020869,
"auc_roc": 0.9821761658031088,
"auc_roc_std": 0.0026392679094843995,
"ece": 0.04494034478956999,
"ece_std": 0.005726424915493296,
"brier_score": 0.0466544096876615,
"brier_score_std": 0.004742563582724417
},
"graph_only": {
"accuracy": 0.9327846364883402,
"accuracy_std": 0.009699681497757861,
"f1_macro": 0.8999754908482798,
"f1_macro_std": 0.015973872212778746,
"f1_weighted": 0.9336681613632254,
"f1_weighted_std": 0.010016150131483496,
"precision": 0.8112522686025408,
"precision_std": 0.025279588089643226,
"recall": 0.88,
"recall_std": 0.05887840577551894,
"auc_roc": 0.9794473229706391,
"auc_roc_std": 0.004518842407326572,
"ece": 0.04988376339660561,
"ece_std": 0.005270687352787256,
"brier_score": 0.05351088606272011,
"brier_score_std": 0.004712551086056143
},
"clusters_only": {
"accuracy": 0.9327846364883401,
"accuracy_std": 0.005132589007920335,
"f1_macro": 0.899547119319776,
"f1_macro_std": 0.008336212842531497,
"f1_weighted": 0.9335092210152506,
"f1_weighted_std": 0.0050248084073146555,
"precision": 0.8196119582735597,
"precision_std": 0.04795456749007511,
"recall": 0.8733333333333334,
"recall_std": 0.06798692684790376,
"auc_roc": 0.9778238341968913,
"auc_roc_std": 0.0032976605139516603,
"ece": 0.04968188728170691,
"ece_std": 0.00536472638852737,
"brier_score": 0.051895951665629936,
"brier_score_std": 0.0025601528689546227
},
"all_spatial": {
"accuracy": 0.9368998628257886,
"accuracy_std": 0.008455986286651548,
"f1_macro": 0.9045052333319902,
"f1_macro_std": 0.01712508160361251,
"f1_weighted": 0.9372024686824588,
"f1_weighted_std": 0.00980954743271344,
"precision": 0.8317384370015949,
"precision_std": 0.02366509448597023,
"recall": 0.8733333333333334,
"recall_std": 0.08055363982396381,
"auc_roc": 0.9836614853195164,
"auc_roc_std": 0.004274602676148179,
"ece": 0.0495688676028967,
"ece_std": 0.0010367371583657778,
"brier_score": 0.047661818741178,
"brier_score_std": 0.004223029229299427
}
},
"clustering_granularity": {
"k=5": {
"accuracy": 0.934156378600823,
"accuracy_std": 0.008889904936087593,
"f1_macro": 0.9001535749642354,
"f1_macro_std": 0.01891210759908036,
"f1_weighted": 0.9343987099876888,
"f1_weighted_std": 0.010627995331271404,
"precision": 0.8259424959083086,
"precision_std": 0.024448942611041833,
"recall": 0.8666666666666667,
"recall_std": 0.08993825042154695,
"auc_roc": 0.9837996545768566,
"auc_roc_std": 0.00397791907443432,
"ece": 0.04689666052261382,
"ece_std": 0.009314414641981877,
"brier_score": 0.04679048982498124,
"brier_score_std": 0.004739284217247446
},
"k=10": {
"accuracy": 0.934156378600823,
"accuracy_std": 0.0058198088986546855,
"f1_macro": 0.9011361082078723,
"f1_macro_std": 0.012701781232144856,
"f1_weighted": 0.9347346656359964,
"f1_weighted_std": 0.007051327186642521,
"precision": 0.8206203915586067,
"precision_std": 0.019237331702761415,
"recall": 0.8733333333333334,
"recall_std": 0.06599663291074444,
"auc_roc": 0.982832469775475,
"auc_roc_std": 0.00368260601169675,
"ece": 0.04728646662768171,
"ece_std": 0.0005582092082208497,
"brier_score": 0.047569526926073726,
"brier_score_std": 0.004569314921970368
},
"k=20": {
"accuracy": 0.935528120713306,
"accuracy_std": 0.006994539799167062,
"f1_macro": 0.9029439707642372,
"f1_macro_std": 0.01415438702923127,
"f1_weighted": 0.9360152806169815,
"f1_weighted_std": 0.00810996688307506,
"precision": 0.825424067414914,
"precision_std": 0.016897875380354546,
"recall": 0.8733333333333334,
"recall_std": 0.06599663291074444,
"auc_roc": 0.983419689119171,
"auc_roc_std": 0.0037228859648518356,
"ece": 0.045144637029069394,
"ece_std": 0.002866283685321084,
"brier_score": 0.046910290821234955,
"brier_score_std": 0.004769634243694431
},
"k=50": {
"accuracy": 0.934156378600823,
"accuracy_std": 0.008889904936087593,
"f1_macro": 0.9008292051083737,
"f1_macro_std": 0.01796672639626433,
"f1_weighted": 0.9346282380308591,
"f1_weighted_std": 0.010307758547442316,
"precision": 0.8206022719634335,
"precision_std": 0.0169135960243027,
"recall": 0.8733333333333334,
"recall_std": 0.08055363982396381,
"auc_roc": 0.9835924006908462,
"auc_roc_std": 0.0032904160840371978,
"ece": 0.0470419108293246,
"ece_std": 0.0013514542902060575,
"brier_score": 0.04743248205297146,
"brier_score_std": 0.003762859331949097
},
"all_granularities": {
"accuracy": 0.9382716049382717,
"accuracy_std": 0.010080204702811431,
"f1_macro": 0.9069658178347435,
"f1_macro_std": 0.01827636538115004,
"f1_weighted": 0.9387027781479818,
"f1_weighted_std": 0.010990149880983232,
"precision": 0.831376448367295,
"precision_std": 0.017573686066550933,
"recall": 0.88,
"recall_std": 0.07118052168020872,
"auc_roc": 0.9833160621761659,
"auc_roc_std": 0.003856073518366977,
"ece": 0.04826353248959792,
"ece_std": 0.00527439530150194,
"brier_score": 0.04663285117900793,
"brier_score_std": 0.004319834384511459
}
},
"transformer_embeddings": {
"pure_gbdt": {
"accuracy": 0.9368998628257886,
"accuracy_std": 0.008455986286651548,
"f1_macro": 0.9045052333319902,
"f1_macro_std": 0.01712508160361251,
"f1_weighted": 0.9372024686824588,
"f1_weighted_std": 0.00980954743271344,
"precision": 0.8317384370015949,
"precision_std": 0.02366509448597023,
"recall": 0.8733333333333334,
"recall_std": 0.08055363982396381,
"auc_roc": 0.9836614853195164,
"auc_roc_std": 0.004274602676148179,
"ece": 0.0495688676028967,
"ece_std": 0.0010367371583657778,
"brier_score": 0.047661818741178,
"brier_score_std": 0.004223029229299427
},
"hybrid_d32_L2": {
"accuracy": 0.9122085048010974,
"accuracy_std": 0.005132589007920374,
"f1_macro": 0.8679139457693156,
"f1_macro_std": 0.015117909003304474,
"f1_weighted": 0.9128605777221734,
"f1_weighted_std": 0.007487299446980015,
"precision": 0.772779817761458,
"precision_std": 0.02916919077517404,
"recall": 0.82,
"recall_std": 0.0864098759787715,
"auc_roc": 0.965699481865285,
"auc_roc_std": 0.006645603745840797,
"ece": 0.07550869523136981,
"ece_std": 0.0070558961501389,
"brier_score": 0.07550888452830928,
"brier_score_std": 0.005517776288285954
},
"hybrid_d64_L2": {
"accuracy": 0.9067215363511659,
"accuracy_std": 0.0038798725991031234,
"f1_macro": 0.8565453270262061,
"f1_macro_std": 0.0065005817962412804,
"f1_weighted": 0.9063543838672204,
"f1_weighted_std": 0.0022026074905745555,
"precision": 0.7860328239954727,
"precision_std": 0.06494706306162124,
"recall": 0.7733333333333333,
"recall_std": 0.08993825042154693,
"auc_roc": 0.9599654576856649,
"auc_roc_std": 0.007181645380979579,
"ece": 0.0807011822330138,
"ece_std": 0.007084585049665119,
"brier_score": 0.08036028322753526,
"brier_score_std": 0.00522874547188044
},
"hybrid_d128_L2": {
"accuracy": 0.9176954732510287,
"accuracy_std": 0.005819808898654738,
"f1_macro": 0.8762910145445768,
"f1_macro_std": 0.007222245542103157,
"f1_weighted": 0.9183569294926975,
"f1_weighted_std": 0.004886826202400205,
"precision": 0.7912580837284336,
"precision_std": 0.050501112169752636,
"recall": 0.8266666666666667,
"recall_std": 0.061824123303304696,
"auc_roc": 0.9654231433506045,
"auc_roc_std": 0.0067817392941276105,
"ece": 0.07278331326166544,
"ece_std": 0.0036995712404727364,
"brier_score": 0.07185565814209835,
"brier_score_std": 0.001206032336780547
}
},
"calibration": {
"uncalibrated": {
"ECE": 0.03786273886144674,
"AUC": 0.9834427173287277,
"Brier": 0.04766181874117801
},
"platt": {
"ECE": 0.02927860504727985,
"AUC": 0.9679332181922856,
"Brier": 0.04826784034107657
},
"isotonic": {
"ECE": 0.03399572434860025,
"AUC": 0.976419113413932,
"Brier": 0.04687223107207046
},
"temperature": {
"ECE": 0.3911245326292897,
"AUC": 0.9793091537132989,
"Brier": 0.22891487324779877
}
}
}