Place: Departamento de Ciência de Computadores, Sala S3

Abstract:

Ubiquitous data mining deals with the problem of data analysis in mobile and distributed environments that we deal with at every step of life. This talk will offer a perspective of ubiquitous data mining by tracing the history and following the current emergence of commercial applications along with future research issues. This talk will particularly explore some of the research challenges in the context of ubiquitous data stream mining in vehicles. Vehicles (e.g. cars, light and heavy-duty trucks) generate data streams of hundreds of parameters from different vehicle sub-systems. These data streams provide wealth of information which can be used to monitor vehicle-health, fuel consumption characteristics, driver behavior and greenhouse emission. Analyzing the data on board the vehicle has a strong business motivation. This talk will describe the MineFleet, a commercial ubiquitous data mining system for vehicles. The talk will end with a discussion on the research challenges for the next generation of ubiquitous data mining research.