{"id":"W3122916539","doi":"","title":"Mind the Gap: Accounting for Measurement Error and Misclassification in Variables Generated via Data Mining","year":2017,"lang":"en","type":"article","venue":"SSRN Electronic Journal","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Quest University Canada","funders":"","keywords":"Econometrics; Computer science; Econometric model; Covariate; Data mining; Errors-in-variables models; Observational error; Variance (accounting); Inference; Parameterized complexity; Statistics; Machine learning; Artificial intelligence; Algorithm; Mathematics; Accounting; Economics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00593477,0.0001173647,0.0001442236,0.00009581239,0.0007859765,0.0005400638,0.002132235,0.00004674587,9.413822e-7],"category_scores_gemma":[0.0004022519,0.0000863484,0.00002937241,0.0001119382,0.0000476669,0.001145112,0.0002975048,0.0004716763,6.61592e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003804709,"about_ca_system_score_gemma":0.0006556287,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004412748,"about_ca_topic_score_gemma":0.002067423,"domain_scores_codex":[0.9981167,0.00006807711,0.00028701,0.0003466241,0.0002908826,0.0008907133],"domain_scores_gemma":[0.998279,0.00005914007,0.0004522393,0.001001219,0.0001775758,0.00003078375],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002257949,0.0000634247,0.003237708,0.00000804873,0.0001723444,0.000002087785,0.0003560582,0.00009682109,0.03168454,0.04300464,0.0001845239,0.9211673],"study_design_scores_gemma":[0.00107874,0.0001393734,0.004386454,0.00009431359,0.00009951375,0.0001966471,0.0005310955,0.7380668,0.004209314,0.2464449,0.004324327,0.0004284997],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01738305,0.001204186,0.9788839,0.002210608,0.00005770459,0.0001663224,9.053634e-7,0.00001841686,0.00007487711],"genre_scores_gemma":[0.937748,0.000339712,0.06165954,0.00003850937,0.0001161027,0.0000141157,0.000003213998,0.000009847848,0.0000709154],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9207388,"threshold_uncertainty_score":0.6045175,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09644917981351636,"score_gpt":0.331303410127973,"score_spread":0.2348542303144567,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}