--- AA1-2012 CUP --- 11 Dec 2012 data sets v1 Information: -) See Slides of AA1-2012 Lecture 17 (motivation, input format, report hints and rules) -) In particular consider that the evaluation is computed according to the Mean Euclidean Error MEE = (1/N) * sum_p^N(sqrt((output_p,x - target_p,x)^2 + (output_p,y -target_p,y)^2)), where N is the number of patterns, p the pattern. -) Data includes Training set (LOC-TR.csv) and corresponding Blind Test Set (LOC-TS.csv) Input format: see file prologue. -) Submission Rules: - Participation is initially restricted to students of AA1 course 2012 - University of Pisa. All participants with their submission agree to the disclosure and publication of their name, results with preliminary and final ranking. - Prepare 1 file in standard text format (comma separated values) team-name_LOC-TS.csv Format of team-name_LOC-TS.csv file: # name of the candidates # nickname of the team (max 8-10 char) for the results page on the web # name of the data set (LOC) - AA1 2012 CUP v1 # date Table with 136 rows, 3 columns (comma separated values) Each row is: id, output_x, output_y (outputs are real valued localization estimations on the X and Y axis, respectively) - Prepare Code package and Report (see Lecture 17 slides) - Prepare short abstract (5 lines) with method and validation descriptions: file name: team-name_abstract.txt - Send 1 file csv and 1 file txt (plus Code package and Report) to: micheli@di.unipi.it - Deadline (and final results): September 2013 --- The competition is conducted purely for academic purposes and should not be exploited commercially. AA1-2012 CUP CREDITS: Computational Intelligence and Machine Learning (CIML) group - Pisa Claudio Gallicchio, Alessio Micheli 2012 (co-starring 2012: P.B. & S.C.) Universita` di Pisa - Dipartimento di Informatica Largo Bruno Pontecorvo 3, 56127 Pisa, ITALY --- Other info: http://pages.di.unipi.it/micheli/DID/CUP-AA1/ Alessio Micheli, Email:micheli@di.unipi.it 11 Dec 2012 ------------------------------------------