Skip to main content

Advertisement

Table 2 Final results for the classification of mystery samples from mystery set 1 and 3

From: Identification of city specific important bacterial signature for the MetaSUB CAMDA challenge microbiome data

SampleRandom ForestSVMFinal PredictionReal LabelStatus of the prediction
PredictionScoreDeparturesAdj. ScorePredictionScoreDeparturesAdj. Score
Mystery Set 1
 C1.001SAC1.00001.000SAC0.84820.240SACSCL 
 C1.002SCL1.00001.000SCL1.00001.000SCLSCLCORRECT
 C1.003NYC1.00001.000NYC0.77110.297NYCOFA 
 C1.004PXO1.00001.000PXO1.00010.500PXOPXOCORRECT
 C1.005NYC1.00001.000OFA1.00010.500NYCOFA 
 C1.006PXO0.99910.499PXO0.82130.168PXOPXOCORRECT
 C1.007SCL0.97110.471SCL0.76910.296SCLSCLCORRECT
 C1.008PXO1.00001.000PXO0.69630.121PXOPXOCORRECT
 C1.009NYC1.00001.000OFA0.61910.192NYCNYCCORRECT
 C1.010PXO1.00001.000PXO0.69820.162PXOPXOCORRECT
 C1.011SCL1.00001.000SCL0.74140.110SCLSCLCORRECT
 C1.012OFA1.00001.000OFA1.00001.000OFAOFACORRECT
 C1.013PXO1.00001.000PXO0.86420.249PXOPXOCORRECT
 C1.014SAC1.00001.000SCL0.71720.171SACSCL 
 C1.015TOK1.00001.000HAM0.46230.053TOKNYC 
 C1.016OFA0.91310.416NYC0.82610.341OFANYC 
 C1.017SCL0.61010.186TOK0.54330.074SCLPXO 
 C1.018NYC1.00001.000NYC0.99510.495NYCNYCCORRECT
 C1.019AKL1.00001.000OFA1.00001.000InconclusiveNYC 
 C1.020OFA1.00001.000OFA1.00010.500OFAOFACORRECT
 C1.021AKL0.83430.174OFA0.99730.248OFANYC 
 C1.022PXO1.00001.000PXO0.89410.399PXOPXOCORRECT
 C1.023NYC1.00010.500NYC0.99020.327NYCNYCCORRECT
 C1.024NYC0.85210.363NYC0.89840.161NYCNYCCORRECT
 C1.025NYC1.00001.000NYC0.99740.199NYCNYCCORRECT
 C1.026PXO1.00001.000PXO1.00001.000PXOPXOCORRECT
 C1.027PXO1.00001.000TOK0.62110.193PXOPXOCORRECT
 C1.028OFA1.00001.000OFA1.00001.000OFAOFACORRECT
 C1.029AKL1.00010.500NYC0.49430.061AKLNYC 
 C1.030TOK0.76110.290TOK0.99410.494TOKPXO 
Mystery Set 3
 C5.001Boston1.00001.000Boston1.00001.000BostonBostonCORRECT
 C5.002Ilorin1.00001.000Lisbon0.75410.284IlorinIlorinCORRECT
 C5.003Lisbon1.00001.000Lisbon1.00001.000LisbonLisbonCORRECT
 C5.004Ilorin1.00001.000Ilorin0.56810.161IlorinIlorinCORRECT
 C5.005Lisbon1.00001.000Lisbon0.99910.499LisbonLisbonCORRECT
 C5.006Lisbon1.00001.000Ilorin0.61610.190LisbonIlorin 
 C5.007Boston1.00001.000Lisbon0.74910.280BostonBogota 
 C5.008Lisbon0.99910.499Lisbon0.77210.298LisbonBogota 
 C5.009Lisbon1.00001.000Lisbon1.00001.000LisbonLisbonCORRECT
 C5.010Ilorin0.38420.049Boston0.98210.482BostonBogota 
 C5.011Lisbon1.00001.000Lisbon1.00001.000LisbonBogota 
 C5.012Lisbon1.00001.000Lisbon0.98810.488LisbonLisbonCORRECT
 C5.013Boston1.00001.000Boston0.99810.498BostonBostonCORRECT
 C5.014Ilorin1.00001.000Ilorin1.00001.000IlorinIlorinCORRECT
 C5.015Boston1.00001.000Lisbon0.75010.282BostonBostonCORRECT
 C5.016Boston0.84310.356Lisbon0.75010.282BostonBogota 
  1. Table shows samples abbreviated names, partial results from both classifiers (RF and SVM) and voted results, actual label of each sample, and whether the samples prediction was correct. Results for sample C1.019 were not correct but also labeled as inconclusive since both classifiers predicted a different city with the same adjusted score. Additionally, in similar cases whether or not one of the classifiers was correct or not was irrelevant due to the inability of the pipeline to produce a label