Web scraping with R & novel classification algorithms on unbalanced data

This is the schedule of our next RBelgium meetup held on Wednesday February 24, 2016 at Paleizenstraat 153, Brussels.

• 18h00-18h30: enter & meet other R users

• 18h30-19h00: Web scraping with R: live scraping products & prices of http://www.delhaize.be

http://photos1.meetupstatic.com/photos/event/e/9/7/e/600_446999774.jpeg

• 19h15-20h00: State-of-the-art classification algorithms with unbalanced data. Package unbalanced: Racing for Unbalanced Methods Selection. (Andrea Dal Pozzolo).

http://photos3.meetupstatic.com/photos/event/e/9/b/5/600_446999829.jpeg Abstract:

State-of-the-art classification algorithms suffer when the data is skewed towards one class. This led to the development of a number of techniques to cope with unbalanced data. However, no technique appears to work consistently better in all conditions. This talk presents a new R package, called unbalanced, which implements some well-known techniques for unbalanced classification tasks and provides a racing strategy to adaptively select the best methods for a given dataset, classification algorithms and accuracy measure adopted.

Bio:

Andrea Dal Pozzolo is a PhD candidate at the Machine Learning Group (MLG) of Université Libre de Bruxelles (ULB), Brussels (Belgium). His research focuses on Machine Learning techniques for Fraud Detection in electronic transactions. In particular he is interested in techniques for unbalanced data streams. He graduated in 2011 with summa cum laude from a master degree from the faculty of Statistics, University of Bologna (Italy).