{"id":"2dfbd9d0-7ffc-41d5-a4a1-451a11995799","arxiv_id":"1908.00417","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A single NCD-plus-SVM pipeline reaches 0.77 to 0.95 accuracy on five text classification tasks, but the parameter-free claim is undercut by per-dataset tuning and training-set evaluation.","lead":"The authors test one compression-based text representation, fed into a single classifier, across five cybersecurity text tasks including malicious URLs, spam, and fake domains. The potential payoff is a unified, low-configuration anomaly detector, but the advertised parameter-free status is not supported by the paper's own tuning choices.","discovery_kind":"extension","skeptic_critique":null,"referee_report":null,"author_rebuttal":null,"desk_editor":null,"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-14T15:58:06.452928+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}