{"id":"W2358452462","doi":"10.1088/1755-1315/34/1/012006","title":"Acquiring geographical data with web harvesting","year":2016,"lang":"en","type":"article","venue":"IOP Conference Series Earth and Environmental Science","topic":"Geographic Information Systems Studies","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Nova Scotia Community College","funders":"","keywords":"Computer science; World Wide Web; Python (programming language); Web page; Database; Information retrieval; Table (database)","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":["sts"],"consensus_categories":["sts"],"category_scores_codex":[0.0009347516,0.0001245177,0.0001263824,0.0000994418,0.001622366,0.000266882,0.0005788306,0.00003909616,0.000162211],"category_scores_gemma":[0.0001050811,0.00007937667,0.00001282993,0.0003364235,0.006917699,0.003175356,0.0004664442,0.00006029192,0.00003296821],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000192085,"about_ca_system_score_gemma":0.0001280586,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004511907,"about_ca_topic_score_gemma":0.002570091,"domain_scores_codex":[0.9982064,0.00003738172,0.0001770118,0.0003943837,0.0006804601,0.0005043158],"domain_scores_gemma":[0.9992406,0.00005870488,0.00009282037,0.0003431462,0.0000266242,0.0002381324],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00009349817,0.0000666189,0.60671,0.00003506076,0.00004841357,0.00001725626,0.02019903,0.00000398321,0.07748687,0.0686355,0.00008010786,0.2266236],"study_design_scores_gemma":[0.001137201,0.000367397,0.7391259,0.0003144611,0.00002857648,0.00007786204,0.06687525,0.0001522418,0.004739799,0.0005588913,0.1856351,0.0009872708],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9797996,0.00005893676,0.0001377446,0.001359687,0.0001005749,0.0001676179,0.00004232428,0.00006192865,0.01827155],"genre_scores_gemma":[0.9976301,0.0005494584,0.0006423691,0.0000660154,0.00004423613,0.000008128995,0.00000339474,0.000003749639,0.001052571],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2256363,"threshold_uncertainty_score":0.9996774,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02841164597280102,"score_gpt":0.2414243143527749,"score_spread":0.2130126683799739,"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."}}