{"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/human-guided-data-exploration-using","title":"Human-guided data exploration using randomisation","arxiv_id":"1805.07725","date":"2018-05-20","proceeding":null,"authors":["Kai Puolamäki","Emilia Oikarinen","Buse Atli","Andreas Henelius"],"abstract":"An explorative data analysis system should be aware of what the user already\nknows and what the user wants to know of the data: otherwise the system cannot\nprovide the user with the most informative and useful views of the data. We\npropose a principled way to do exploratory data analysis, where the user's\nbackground knowledge is modeled by a distribution parametrised by subsets of\nrows and columns in the data, called tiles. The user can also use tiles to\ndescribe his or her interests concerning relations in the data. We provide a\ncomputationally efficient implementation of this concept based on constrained\nrandomisation. The implementation is used to model both the background\nknowledge and the user's information request and is a necessary prerequisite\nfor any interactive system. Furthermore, we describe a novel linear projection\npursuit method to find and show the views most informative to the user, which\nat the limit of no background knowledge and with generic objectives reduces to\nPCA. We show that our method is robust under noise and fast enough for\ninteractive use. We also show that the method gives understandable and useful\nresults when analysing real-world data sets. We will release an open source\nlibrary implementing the idea, including the experiments presented in this\npaper. We show that our method can outperform standard projection pursuit\nvisualisation methods in exploration tasks. Our framework makes it possible to\nconstruct human-guided data exploration systems which are fast, powerful, and\ngive results that are easy to comprehend.","url_abs":"http://arxiv.org/abs/1805.07725v2","url_pdf":"http://arxiv.org/pdf/1805.07725v2.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":"human-guided-data-exploration-using","repo_url":"https://github.com/edahelsinki/corand","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}