BEGIN:VCALENDAR PRODID;X-RICAL-TZSOURCE=TZINFO:-//Calagator//EN CALSCALE:GREGORIAN X-WR-CALNAME:Calagator METHOD:PUBLISH VERSION:2.0 BEGIN:VTIMEZONE TZID;X-RICAL-TZSOURCE=TZINFO:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20120311T020000 RDATE:20120311T020000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 TZNAME:PDT END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT CREATED;VALUE=DATE-TIME:20120504T212553Z DTEND;TZID=America/Los_Angeles;VALUE=DATE-TIME:20120511T113000 DTSTART;TZID=America/Los_Angeles;VALUE=DATE-TIME:20120511T103000 DTSTAMP;VALUE=DATE-TIME:20120504T212553Z LAST-MODIFIED;VALUE=DATE-TIME:20120504T212553Z UID:http://calagator.org/events/1250462359 DESCRIPTION:Presented by Charles Parker \;\n \;\nA basic problem in computer science is binary classification\, in which an algorithm appli es a binary label to data based on the presence or absence of some pheno menon. Problems of this type abound in areas as diverse as computational biology\, multimedia indexing\, and anomaly detection. Evaluating the p erformance of a binary labeling algorithm is itself a complex task\, oft en based on a domain-dependent notion of the relative cost of "\;fal se positives"\; versus "\;false negatives"\;. As these costs are often not available to researchers or engineers\, a number of metho ds are used to provide a cost-independent analysis of performance. In th is talk\, I will examine a number of these methods both theoretically an d experimentally. The presented results suggest a set of best practices for evaluating binary classification algorithms\, while questioning whet her a cost-independent analysis is even possible. \n\nTags: galois\, ana lysis\, tech talk\, algorithms\n\nImported from: http://calagator.org/ev ents/1250462359 URL:http://corp.galois.com/blog/2012/5/4/tech-talk-an-analysis-of-analysi s.html SUMMARY:Galois Tech Talk: An Analysis of Analysis LOCATION:Galois\, Inc: 421 SW 6th Ave. Suite 300\, Portland OR 97204 US SEQUENCE:1 END:VEVENT END:VCALENDAR