{"id":"W2137763598","doi":"10.1109/tkde.2004.58","title":"Efficient phrase-based document indexing for Web document clustering","year":2004,"lang":"en","type":"article","venue":"IEEE Transactions on Knowledge and Data Engineering","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":320,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Document clustering; Computer science; Cluster analysis; Vector space model; Search engine indexing; Phrase; Information retrieval; Document classification; Data mining; tf–idf; Artificial intelligence; Term (time)","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":[0.001256536,0.000784139,0.001399069,0.004405822,0.001098526,0.00160535,0.001782839,0.001007364,0.002336452],"category_scores_gemma":[0.007110163,0.0004591109,0.001024594,0.009499787,0.0006611206,0.002796623,0.001643425,0.001270631,0.002618118],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001103874,"about_ca_system_score_gemma":0.001569829,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004000303,"about_ca_topic_score_gemma":0.004178685,"domain_scores_codex":[0.9983574,0.0004335535,0.0001242489,0.0002648279,0.0007371228,0.00008283414],"domain_scores_gemma":[0.9980203,0.0006195031,0.0001676637,0.0005740468,0.0005630313,0.00005553494],"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":[0.0002013628,0.0001207917,0.0008992539,0.0004222789,0.0001033649,0.0001075881,0.0002800754,0.0887168,0.02255181,0.05916213,0.01699824,0.8104363],"study_design_scores_gemma":[0.00004592617,0.00008310653,0.0005826834,0.00002482439,0.0000490034,0.0002336378,0.00007138492,0.9032925,0.008853515,0.07417627,0.01253762,0.00004959584],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003313171,0.0003755327,0.993883,0.0000627565,0.00003433012,0.00009091949,0.0003639852,0.00116452,0.0007118771],"genre_scores_gemma":[0.03637255,0.0005158997,0.9601054,0.00004729038,0.00007315654,0.0002402213,0.001676647,0.00021147,0.0007574298],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004405822,"threshold_uncertainty_score":0.008009255,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02028388213544066,"score_gpt":0.2679597706680517,"score_spread":0.247675888532611,"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."}}