{"id":"W1595390009","doi":"","title":"Cybergenre: automatic identification of home pages on the web","year":2004,"lang":"en","type":"article","venue":"Journal of Web Engineering","topic":"Authorship Attribution and Profiling","field":"Computer Science","cited_by":40,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Web page; Home page; Computer science; World Wide Web; Classifier (UML); Information retrieval; Static web page; Web development; The Internet; Artificial intelligence","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":[0.0006991696,0.0004712734,0.0004645302,0.00547107,0.0003238225,0.001083506,0.0003929017,0.000555839,0.002562352],"category_scores_gemma":[0.002589651,0.0001480045,0.0002600815,0.001659679,0.0002016907,0.001966214,0.0006500983,0.0003932257,0.002060993],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002910169,"about_ca_system_score_gemma":0.0002781237,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001675951,"about_ca_topic_score_gemma":0.00305488,"domain_scores_codex":[0.9994608,0.0001415848,0.00004599652,0.0001287626,0.0001592258,0.00006368572],"domain_scores_gemma":[0.9980952,0.0006870924,0.0003027863,0.0003786519,0.0003919767,0.0001443538],"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":[0.0008828549,0.0005534642,0.1017389,0.000482791,0.0001170419,0.0005770517,0.0006385991,0.004080236,0.03143503,0.001799649,0.01772431,0.8399702],"study_design_scores_gemma":[0.00008390744,0.0004569212,0.2794515,0.0001743153,0.0001130429,0.002444292,0.001547172,0.5982065,0.08206464,0.006963381,0.02838984,0.000104475],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8749635,0.001252102,0.08725727,0.0002586371,0.0001400272,0.0003959782,0.006961149,0.01734345,0.0114278],"genre_scores_gemma":[0.8783457,0.0004575335,0.1015119,0.0000532489,0.00007881867,0.0001156066,0.01053847,0.0002402934,0.008658449],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.00547107,"threshold_uncertainty_score":0.008571923,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01742234240319396,"score_gpt":0.2319652182478575,"score_spread":0.2145428758446635,"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."}}