{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001236695,0.0007280611,0.0007033407,0.007674752,0.000937424,0.002738481,0.0009874589,0.0006526103,0.009722263],"category_scores_gemma":[0.007157196,0.00072158,0.001023632,0.0102051,0.0005176271,0.002882483,0.00332435,0.0008672878,0.009279566],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006561999,"about_ca_system_score_gemma":0.001214474,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01035451,"about_ca_topic_score_gemma":0.01149481,"domain_scores_codex":[0.9985479,0.000207724,0.0001581781,0.0003100355,0.0006975146,0.00007866018],"domain_scores_gemma":[0.9954462,0.001091924,0.0002681854,0.001869466,0.001162573,0.0001616403],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002069321,0.0002809742,0.02720759,0.0008826834,0.0001976361,0.001159039,0.002472265,0.01264128,0.02800939,0.008136264,0.05787158,0.8609343],"study_design_scores_gemma":[0.000114838,0.0002482048,0.1367695,0.000633479,0.0002675892,0.002401869,0.009421741,0.1536436,0.08304086,0.07507758,0.5379058,0.0004750467],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06652489,0.0006396833,0.7903101,0.0008944951,0.0002440917,0.001418369,0.04353559,0.05721696,0.03921576],"genre_scores_gemma":[0.1769855,0.001034687,0.7580478,0.0002952028,0.0001091399,0.0007064316,0.05052012,0.002907457,0.009393612],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01035451,"threshold_uncertainty_score":0.03252423,"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."}}