{"id":"W2032328503","doi":"10.1007/s10791-009-9108-x","title":"Document clustering of scientific texts using citation contexts","year":2009,"lang":"en","type":"article","venue":"Information Retrieval","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":77,"is_retracted":false,"has_abstract":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Cluster analysis; Information retrieval; Computer science; Document clustering; Citation; Vocabulary; Context (archaeology); Similarity (geometry); Representation (politics); Document retrieval; Natural language processing; Artificial intelligence; World Wide Web; Linguistics","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":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.001669988,0.0008086043,0.001106689,0.02648126,0.002412321,0.004088819,0.001049552,0.001127128,0.003157671],"category_scores_gemma":[0.01184732,0.0003891796,0.001247355,0.02233986,0.0006420256,0.002487009,0.00120861,0.0009756474,0.002097595],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001043688,"about_ca_system_score_gemma":0.002374801,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00395782,"about_ca_topic_score_gemma":0.008434736,"domain_scores_codex":[0.9978617,0.0005519332,0.0003084471,0.0004282437,0.0006915503,0.0001581303],"domain_scores_gemma":[0.9923053,0.003579697,0.0006003455,0.0005478903,0.002680715,0.0002859683],"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":[0.0008256128,0.0004109048,0.01502229,0.001657673,0.0004256607,0.0003710898,0.001342308,0.01013203,0.03755221,0.01368703,0.01772108,0.900852],"study_design_scores_gemma":[0.000441689,0.001151652,0.100115,0.001224836,0.003434505,0.002277089,0.003867063,0.507206,0.1036586,0.1263126,0.1497516,0.0005594292],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4352847,0.02988485,0.488127,0.002138877,0.00194183,0.001189411,0.009118278,0.005384562,0.02693051],"genre_scores_gemma":[0.6016658,0.006460908,0.3663079,0.0001524675,0.001650409,0.0005931752,0.01205742,0.0006688847,0.01044304],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9735187,"threshold_uncertainty_score":0.01056349,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0160256859248469,"score_gpt":0.2950436954188471,"score_spread":0.2790180094940002,"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."}}