{"id":"W4312729483","doi":"10.2196/40403","title":"Development of a COVID-19–Related Anti-Asian Tweet Data Set: Quantitative Study","year":2022,"lang":"en","type":"article","venue":"JMIR Formative Research","topic":"Data-Driven Disease Surveillance","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Western University; Institute for Christian Studies; University of Toronto","funders":"","keywords":"Social media; Set (abstract data type); Computer science; Stigma (botany); Data set; Coronavirus disease 2019 (COVID-19); Information retrieval; Internet privacy; Artificial intelligence; Psychology; World Wide Web; Medicine","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"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":[0.005976222,0.0001966326,0.0005372188,0.0007391213,0.0007455542,0.00002989142,0.001100965,0.00004194474,0.001488384],"category_scores_gemma":[0.0009531469,0.0001769184,0.00005725252,0.00193553,0.0003291003,0.0003927759,0.003129309,0.0009017043,0.0002200245],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006416939,"about_ca_system_score_gemma":0.002618816,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009622128,"about_ca_topic_score_gemma":0.00007919303,"domain_scores_codex":[0.9941318,0.001515166,0.0007735157,0.0005740239,0.002399509,0.000605994],"domain_scores_gemma":[0.997033,0.000447187,0.0002326319,0.001430715,0.0003847526,0.0004717178],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.003903915,0.01314598,0.1341173,0.002042688,0.002725927,0.001076179,0.7073174,0.00008144574,0.00214517,0.00212574,0.1173329,0.01398538],"study_design_scores_gemma":[0.00851614,0.003265746,0.4667682,0.0001112586,0.00004896619,0.00006020532,0.3698577,0.002661895,0.0002210141,0.0002156444,0.1478138,0.000459398],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9890395,0.0001631175,0.0001590465,0.0009795582,0.00008867257,0.002970924,0.003097405,0.0001055525,0.003396263],"genre_scores_gemma":[0.9948924,0.000009150353,0.0009542124,0.00007581638,0.00001161774,0.0004638979,0.003311674,0.00003226995,0.0002490282],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3374597,"threshold_uncertainty_score":0.9994244,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2234980809995228,"score_gpt":0.5090776074569722,"score_spread":0.2855795264574494,"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."}}