{"id":"W2293475867","doi":"","title":"Towards web search engine scale data mining","year":2009,"lang":"en","type":"article","venue":"Australasian Data Mining Conference","topic":"Data Management and Algorithms","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Scalability; Adaptability; Search engine; Web mining; Data science; Data mining; Context (archaeology); Web search query; Web crawler; Web search engine; Metasearch engine; Search analytics; Construct (python library); Big data; Information retrieval; World Wide Web; Database; Web service","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":[0.01003651,0.001220451,0.002984072,0.004561246,0.0009197942,0.004947164,0.002764899,0.002516437,0.001040821],"category_scores_gemma":[0.04024334,0.001583882,0.002387036,0.006794104,0.001662759,0.009602892,0.003852092,0.006941668,0.00186048],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001297401,"about_ca_system_score_gemma":0.002027587,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002675886,"about_ca_topic_score_gemma":0.002986814,"domain_scores_codex":[0.9940436,0.002425056,0.0004136929,0.001067979,0.001860364,0.0001893493],"domain_scores_gemma":[0.9732766,0.01633543,0.001176,0.005349915,0.003195586,0.0006663943],"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":[0.0004594266,0.0014805,0.0293846,0.001874151,0.001035428,0.0009538346,0.001577031,0.2198498,0.01657764,0.2293893,0.042003,0.4554153],"study_design_scores_gemma":[0.00005921153,0.00005522839,0.00126737,0.00008581197,0.00004755408,0.0001505737,0.0001984122,0.7455148,0.002116469,0.2371021,0.01336602,0.00003651146],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02247662,0.004302252,0.9592943,0.008032251,0.0001887101,0.0002464259,0.0008455352,0.002752846,0.001861087],"genre_scores_gemma":[0.1144885,0.003904024,0.875137,0.001634346,0.0007722658,0.0005127362,0.001775442,0.0002788544,0.001496857],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01003651,"threshold_uncertainty_score":0.05307877,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1295590154397186,"score_gpt":0.3325322425937539,"score_spread":0.2029732271540353,"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."}}