Table 3

Classification scores and predictions for the 22 subjects with polyps whose home stool samples were called false negative on Naïve Bayes and/or Neural Network models

Sample IDClassClassifierScore HS_Polyp_YScore HS_Polyp_NPredictedCall
A. Subjects where both models call FN
 HS_23HS_Polyp_YNaïve Bayes0.0070.993HS_Polyp_NFN
Neural Network01HS_Polyp_NFN
 HS_341HS_Polyp_YNaïve Bayes0.1590.841HS_Polyp_NFN
Neural Network0.1770.823HS_Polyp_NFN
 HS_372HS_Polyp_YNaïve Bayes0.2130.787HS_Polyp_NFN
Neural Network0.2080.792HS_Polyp_NFN
 HS_381HS_Polyp_YNaïve Bayes0.0050.995HS_Polyp_NFN
Neural Network0.3730.627HS_Polyp_NFN
 HS_384HS_Polyp_YNaïve Bayes0.0260.974HS_Polyp_NFN
Neural Network0.2560.744HS_Polyp_NFN
 HS_386HS_Polyp_YNaïve Bayes0.350.65HS_Polyp_NFN
Neural Network0.4180.582HS_Polyp_NFN
 HS_413HS_Polyp_YNaïve Bayes0.3280.672HS_Polyp_NFN
Neural Network0.4270.573HS_Polyp_NFN
 HS_423HS_Polyp_YNaïve Bayes0.010.99HS_Polyp_NFN
Neural Network0.0020.998HS_Polyp_NFN
 HS_461HS_Polyp_YNaïve Bayes0.5780.422HS_Polyp_NFN
Neural Network0.3030.697HS_Polyp_NFN
B. Subjects where one model calls FN and the other calls TP
 HS_363HS_Polyp_YNaïve Bayes0.7790.221HS_Polyp_YTP
Neural Network0.0340.966HS_Polyp_NFN
 HS_367HS_Polyp_YNaïve Bayes0.9050.095HS_Polyp_YTP
Neural Network0.3650.635HS_Polyp_NFN
 HS_373HS_Polyp_YNaïve Bayes0.5780.422HS_Polyp_NFN
Neural Network0.9920.008HS_Polyp_YTP
 HS_403HS_Polyp_YNaïve Bayes0.6270.373HS_Polyp_NFN
Neural Network0.9820.018HS_Polyp_YTP
 HS_407HS_Polyp_YNaïve Bayes0.0760.924HS_Polyp_NFN
Neural Network0.9910.009HS_Polyp_YTP
 HS_412HS_Polyp_YNaïve Bayes0.3170.683HS_Polyp_NFN
Neural Network0.5280.472HS_Polyp_YTP
 HS_417HS_Polyp_YNaïve Bayes0.8940.106HS_Polyp_YTP
Neural Network0.030.97HS_Polyp_NFN
 HS_420HS_Polyp_YNaïve Bayes0.5240.476HS_Polyp_NFN
Neural Network0.6540.346HS_Polyp_YTP
 HS_427HS_Polyp_YNaïve Bayes0.0840.916HS_Polyp_NFN
Neural Network0.5360.464HS_Polyp_YTP
 HS_45HS_Polyp_YNaïve Bayes0.6510.349HS_Polyp_YTP
Neural Network0.0790.921HS_Polyp_NFN
 HS_507HS_Polyp_YNaïve Bayes0.7110.289HS_Polyp_NFN
Neural Network0.6820.318HS_Polyp_YTP
 HS_6HS_Polyp_YNaïve Bayes0.0310.969HS_Polyp_NFN
Neural Network0.7310.269HS_Polyp_YTP
 HS_62HS_Polyp_YNaïve Bayes0.6860.314HS_Polyp_YTP
Neural Network0.4530.547HS_Polyp_NFN
  • Classification scores are presented for each model and subject where the true class (ie, Polyp-Y or Polyp-N) is tabulated along with the model scores and predictions.

  • FN, false negative;HS, home collected stool sample;Polyp_N, polyp-negative group;Polyp_Y, polyp-positive group;TP, true positive.