{"id":"W3169956799","doi":"10.22215/etd/2021-14417","title":"Search Engines That Scan For Internet-Connected Services: Classification and Empirical Study","year":2021,"lang":"en","type":"dissertation","venue":"","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"","keywords":"The Internet; World Wide Web; Scope (computer science); Computer science; Internet presence management; Internet research; Internet appliance; Search engine; Internet Draft; Internet Connection Sharing; Metasearch engine; Web page; Information retrieval; Internet access; Web search query","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":[0.005467535,0.0005053392,0.0009382988,0.009374368,0.0009773672,0.003786423,0.001190765,0.001374143,0.0037116],"category_scores_gemma":[0.04595596,0.0003187845,0.001078632,0.01543896,0.001409366,0.005810818,0.001453237,0.002314707,0.002364037],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009787621,"about_ca_system_score_gemma":0.001084109,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007035944,"about_ca_topic_score_gemma":0.005638574,"domain_scores_codex":[0.9965674,0.001034715,0.0004445408,0.0003933278,0.001068262,0.000491789],"domain_scores_gemma":[0.9308016,0.04686807,0.01267716,0.002645828,0.004511049,0.002496243],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00009924559,0.0002889601,0.9883342,0.00008580377,0.00008001483,0.00005492048,0.0005970016,0.0001317366,0.00009370298,0.0006027831,0.0008522052,0.008779416],"study_design_scores_gemma":[0.00002623646,0.0001424791,0.9836166,0.0001728956,0.0001702712,0.0007837939,0.00503588,0.006501556,0.0003245821,0.001367758,0.001830938,0.00002690309],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9949422,0.001096033,0.0004537009,0.0005141044,0.00001646123,0.00005118026,0.0009641346,0.00001514364,0.001946891],"genre_scores_gemma":[0.9945103,0.0009506263,0.0005276721,0.0001394176,0.0000947561,0.0000463347,0.003076052,0.00002401134,0.0006308155],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009374368,"threshold_uncertainty_score":0.02891541,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07018848024639512,"score_gpt":0.3505617850748977,"score_spread":0.2803733048285025,"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."}}