{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/explaining-anomalies-in-groups-with","title":"Explaining Anomalies in Groups with Characterizing Subspace Rules","arxiv_id":"1708.05929","date":"2017-08-20","proceeding":null,"authors":["Meghanath Macha","Leman Akoglu"],"abstract":"Anomaly detection has numerous applications and has been studied vastly. We\nconsider a complementary problem that has a much sparser literature: anomaly\ndescription. Interpretation of anomalies is crucial for practitioners for\nsense-making, troubleshooting, and planning actions. To this end, we present a\nnew approach called x-PACS (for eXplaining Patterns of Anomalies with\nCharacterizing Subspaces), which \"reverse-engineers\" the known anomalies by\nidentifying (1) the groups (or patterns) that they form, and (2) the\ncharacterizing subspace and feature rules that separate each anomalous pattern\nfrom normal instances. Explaining anomalies in groups not only saves analyst\ntime and gives insight into various types of anomalies, but also draws\nattention to potentially critical, repeating anomalies.\n  In developing x-PACS, we first construct a desiderata for the anomaly\ndescription problem. From a descriptive data mining perspective, our method\nexhibits five desired properties in our desiderata. Namely, it can unearth\nanomalous patterns (i) of multiple different types, (ii) hidden in arbitrary\nsubspaces of a high dimensional space, (iii) interpretable by the analysts,\n(iv) different from normal patterns of the data, and finally (v) succinct,\nproviding the shortest data description. Furthermore, x-PACS is highly\nparallelizable and scales linearly in terms of data size.\n  No existing work on anomaly description satisfies all of these properties\nsimultaneously. While not our primary goal, the anomalous patterns we find\nserve as interpretable \"signatures\" and can be used for detection. We show the\neffectiveness of x-PACS in explanation as well as detection on real-world\ndatasets as compared to state-of-the-art.","url_abs":"http://arxiv.org/abs/1708.05929v4","url_pdf":"http://arxiv.org/pdf/1708.05929v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"explaining-anomalies-in-groups-with","repo_url":"https://github.com/meghanathmacha/xPACS","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"},{"task_slug":"descriptive","task_name":"Descriptive"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1708.05929","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}