News

27-02-2017 The Let's Roq competition: Check the final results at the Roqad's webpage
27-02-2017 Check the final results of the letsroq competition here. Congratulations to the winners!!!
23-01-2017 Internal submission deadline for OLX and Roq.ad competitions
04-11-2016 The Let's Roq competition: the official kick-off of the competion (Friday, Nov. 4, 9:45 am, room 125 BT)
Talk by Carsten Frien, CEO of Roq.ad
21-10-2016 The Data Ninja competition: the official launch of the competion (Friday, Oct. 21, 9.45 am, room 125 BT
Talk by Tomasz Gramza from OLX
07-10-2016 The challenge has begun :)

The aim and the scope of the challenge

The aim of the course: To learn how to deal with real-life massive data (or to win a data mining competition).

The scope of the course: Application of data mining algorithms to real-life massive data:

Main informations about the course

Time and place

Lectures

Schedule of lectures

07-10-2016 The mining massive data sets challenge [pdf]
14-10-2016 Canceled
21-10-2016 Launching of the "Data Ninja" competition (a talk by Tomasz Gramza fom OLX)
28-10-2016 Canceled
04-11-2016 Launching of the Let's Roq competition (a talk by Carsten Frien, CEO of Roq.ad)
18-11-2016 Presentation of students' reports
25-11-2016 Presentation of students' reports
02-12-2016 Student's talks: Torch and TensorFlow
09-12-2016 Student's talks: Deeplearning4J, GraphLab
16-12-2016 Student's talks: Accord.Net, OpenAi Gym, Theano
13-01-2017 Presentation of students' reports
20-01-2017 Presentation of students' reports

The Challenge

The Data Ninja competition

For more information check: dataninja.olx.pl

Additional information about participation in the competition can be found here

The Let's Roq challenge

For more information check: challenge.roqad.ad

Additional information about participation in the competition can be found here

Reports

Project's meetings

Schedule of the project's meetings

The schedule of the project's meetings can be found here.

Evaluation

Lecture:
Report: data organization/processing: 45 points (min. 50%)
Report: predictive models : 45 points (min. 50%)
Students' talk : 10 points
Top score in the challenge : 10 points

Scale

90% 5.0
80% 4.5
70% 4.0
60% 3.5
50% 3.0