Resources

‘AlgoMost’ Contest for Data Scientists: Forecasting social network dynamic graph

15th Oct `14, 11:55 AM in Resources

International Data Science platform “AlgoMost”, which held a competition early this year among more than 500 data miners…

BDMS
Guest Contributor
 

International Data Science platform “AlgoMost”, which held a competition early this year among more than 500 data miners and data scientists, has launched a new challenge for data scientists: Forecasting development of a social networks’ dynamic graph. The competition will commence on September 17 and will run until November 30, 2014.

With the appearance of social networks dynamic graph has become one of the main subjects of research in data analysis. This graph can change its features in time: its nodes and edges can appear and disappear, a press release from the organizers said.

Importantly, by social networks in mathematics is meant not only Facebook and other online platforms but also the clients of mobile operator (where edges are the facts of calls between subscribers), a scientific research community (where edges are the records of a collaborative study) and etc.

One of the most critical tasks is to forecast the formation or disruption of graph connections. For instance, if some nodes in the graph relate to “services”, “communities” and etc., and the formation of edges between them is interpreted as “service usage” or “joint to community”, the correct forecast of edge’s formation means an opportunity to offer needed services in time, to recommend a network user an appropriate community. In the first case it is possible to charge a fee for “service usage” earlier, in the second – to increase the user’s loyalty by target-oriented recommendations.

To get more information about the content and download data, please follow this link.

MORE FROM BIG DATA MADE SIMPLE