{"id":"W3217062054","doi":"10.3390/math9222988","title":"Can Fake News Detection Models Maintain the Performance through Time? A Longitudinal Evaluation of Twitter Publications","year":2021,"lang":"en","type":"article","venue":"Mathematics","topic":"Misinformation and Its Impacts","field":"Social Sciences","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Robustness (evolution); Computer science; Context (archaeology); Word embedding; Machine learning; Feature engineering; Focus (optics); Artificial intelligence; Longitudinal data; Term (time); Feature (linguistics); Embedding; Data science; Deep learning; Data mining","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":[0.02261654,0.001618626,0.001173567,0.001476556,0.0007929799,0.002505702,0.001264691,0.002386692,0.001193901],"category_scores_gemma":[0.06417777,0.0005672167,0.0009352944,0.0006488416,0.0009358578,0.004909136,0.001338433,0.002493282,0.001563247],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001355879,"about_ca_system_score_gemma":0.0009183878,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007196588,"about_ca_topic_score_gemma":0.005859233,"domain_scores_codex":[0.9928746,0.004080609,0.0005102819,0.001205108,0.0009039672,0.0004255025],"domain_scores_gemma":[0.9479931,0.03345933,0.004054722,0.007602707,0.005471049,0.001419213],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00603587,0.003718555,0.307132,0.0008100882,0.001950244,0.0003571261,0.001566613,0.3408798,0.01467116,0.001549516,0.01233265,0.3089964],"study_design_scores_gemma":[0.00009967716,0.002270313,0.03731759,0.00009212962,0.0002676512,0.0002277333,0.000422467,0.9436401,0.01216906,0.00148661,0.001907756,0.00009899111],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9759244,0.001980479,0.01533696,0.002019875,0.0002482314,0.0001187607,0.001047324,0.000805133,0.002518787],"genre_scores_gemma":[0.9913377,0.000219908,0.005635891,0.0001090842,0.00008040339,0.00004464815,0.001634997,0.00005559955,0.0008818808],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02261654,"threshold_uncertainty_score":0.1196092,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1535578308925872,"score_gpt":0.3610018866829584,"score_spread":0.2074440557903712,"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."}}