{"id":"W2130734227","doi":"10.1109/ccece.2004.1344986","title":"Evaluation of three dimensionality reduction techniques for document classification","year":2004,"lang":"en","type":"article","venue":"","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Dimensionality reduction; Computer science; Curse of dimensionality; Reduction (mathematics); Artificial intelligence; Latent semantic analysis; Categorization; Document classification; Search engine indexing; Machine learning; Data mining; Representation (politics); Measure (data warehouse); Pattern recognition (psychology); Information retrieval; Mathematics","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00162527,0.00007938332,0.00009302572,0.0001243213,0.00008660544,0.00004064674,0.0003596731,0.00007244552,0.0000127888],"category_scores_gemma":[0.0001283392,0.00006761247,0.00005316818,0.0002625624,0.00006032581,0.0005532313,0.00006356077,0.00003895862,0.000006028466],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002298091,"about_ca_system_score_gemma":0.0001534497,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002485776,"about_ca_topic_score_gemma":0.000008534151,"domain_scores_codex":[0.998697,0.00003272858,0.0002866036,0.0002809286,0.0005972643,0.0001054705],"domain_scores_gemma":[0.9985144,0.00002660606,0.0001914452,0.0005391662,0.0007082847,0.00002013267],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000002761627,0.00005400238,0.00001963108,0.000004669103,0.000007518909,8.142579e-9,0.00002666089,0.00003974705,0.02772472,0.7460983,0.0001708995,0.2258511],"study_design_scores_gemma":[0.0002881023,0.00007543992,0.003154186,0.000009299234,0.00001808594,0.000001250105,0.00005379234,0.004045337,0.3394064,0.6524231,0.0004531705,0.00007184686],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02042115,0.00005901874,0.9697827,0.006492405,0.0001450056,0.0008400193,0.000001034114,0.0005441412,0.001714603],"genre_scores_gemma":[0.8066812,0.000008460209,0.1929598,0.00001455083,0.00001484225,0.0002782456,0.000006342457,0.000003005779,0.00003361178],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.78626,"threshold_uncertainty_score":0.2757156,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09872650376107284,"score_gpt":0.3484091910519327,"score_spread":0.2496826872908599,"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."}}