{"id":"W4367839894","doi":"10.36227/techrxiv.22724117","title":"Data-Centric Content Classification of Smart City Residential Services","year":2023,"lang":"en","type":"preprint","venue":"","topic":"Human Mobility and Location-Based Analysis","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Metropolitan area; Context (archaeology); Service (business); Process (computing); Task (project management); Embedding; Segmentation; Data science; Artificial intelligence; Machine learning; Business; Geography; Engineering","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":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002551743,0.0001330068,0.0003195167,0.0002195336,0.0003240703,0.0001523701,0.00140704,0.0002770524,0.001234584],"category_scores_gemma":[0.0003381368,0.0001318889,0.000145488,0.0004588335,0.0002546754,0.0001552492,0.0006493203,0.0002578507,0.0001383383],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000135894,"about_ca_system_score_gemma":0.0006329835,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.2740832,"about_ca_topic_score_gemma":0.5545004,"domain_scores_codex":[0.9973692,0.0004571174,0.0005801152,0.0006302079,0.0007512288,0.0002121622],"domain_scores_gemma":[0.9974936,0.0002582983,0.0004292542,0.001259312,0.0004599811,0.00009961708],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002981479,0.00259893,0.738333,0.005213907,0.002952414,0.00001590958,0.04793002,0.002489349,0.0008171442,0.1083785,0.04633195,0.04464073],"study_design_scores_gemma":[0.000639488,0.00004761925,0.8386078,0.0004964356,0.001361845,1.120377e-7,0.0487684,0.04593286,0.0002865172,0.02230633,0.04049143,0.001061088],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9214139,0.0005862739,0.02222762,0.0198172,0.003226336,0.002589582,0.001879495,0.001004901,0.02725468],"genre_scores_gemma":[0.9916927,0.0003116473,0.0001065315,0.00006385232,0.0003056702,0.00002782541,0.001499692,0.000009855609,0.005982182],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2804172,"threshold_uncertainty_score":0.9996784,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2592285223210208,"score_gpt":0.3829184989199653,"score_spread":0.1236899765989444,"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."}}