{"id":"W4406993411","doi":"10.1007/s10639-025-13370-6","title":"Predicting data science performance from log data: using machine learning","year":2025,"lang":"en","type":"article","venue":"Education and Information Technologies","topic":"Big Data and Business Intelligence","field":"Business, Management and Accounting","cited_by":3,"is_retracted":false,"has_abstract":false,"ca_institutions":"Simon Fraser University","funders":"British Columbia Knowledge Development Fund; Canada Foundation for Innovation","keywords":"Computer science; Educational technology; Science education; Data science; Machine learning; Mathematics education; Artificial intelligence; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0004628836,0.0001074442,0.00009512895,0.0006213312,0.0006025731,0.0007801707,0.001421773,0.00006842128,0.00003123764],"category_scores_gemma":[0.001343401,0.00009541297,0.000005228535,0.001217571,0.0002631008,0.03112147,0.002435941,0.0001914917,0.00004452994],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002807966,"about_ca_system_score_gemma":0.0001837199,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008732862,"about_ca_topic_score_gemma":0.00002180865,"domain_scores_codex":[0.999129,0.000002324523,0.0002766192,0.0002560084,0.000176597,0.0001594654],"domain_scores_gemma":[0.9986714,0.00002381225,0.0002232686,0.0008659303,0.0002109314,0.000004633926],"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":[0.000009332147,0.00003711278,0.2195847,0.0002079488,0.000009565826,4.227392e-8,0.00006678895,0.00006188032,0.0002333809,0.0140864,0.00220711,0.7634957],"study_design_scores_gemma":[0.00009581625,0.000002581816,0.01742172,0.0001724735,0.0000263592,0.000001425998,0.003655515,0.5848116,0.000711051,0.0006715088,0.3922723,0.0001576527],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9646907,0.001281925,0.01088739,0.003452516,0.002118575,0.0003383028,0.00008751841,0.001312521,0.0158305],"genre_scores_gemma":[0.9948087,0.0004940222,0.002327045,0.000553638,0.0001076168,0.000005635719,0.001646187,0.000003474898,0.00005366123],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7633381,"threshold_uncertainty_score":0.9824297,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07699128705415802,"score_gpt":0.3200711976206982,"score_spread":0.2430799105665402,"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."}}