SkyNet - proposition of scalable and distributed machine learning algorithm for BigData problems

Development in hardware, cloud computing and dissemination of the Internet during last decade gave computer scientists power to collect data in the amounts that were not even dreamt of before. Scale of these data sources makes classical relational databases inefficient or useless in storing and processing tasks -- new algorithms had to be invented. The term "Big data" has been created to describe that shift in paradigm.Although there are tools allowing user to store and compute aggregations on Big data, there is still a lot of work to do in order to enable machine learning capabilities. In my thesis I would like to propose a suitable computational model, analyze its advantages over existing models and check its implementation on real-life data.Development in hardware, cloud computing and dissemination of the Internet during last decade gave computer scientists power to collect data in the amounts that were not even dreamt of before. Scale of these data sources makes classical relational databases inefficient or useless in storing and processing tasks -- new algorithms had to be invented. The term "Big data" has been created to describe that shift in paradigm.Although there are tools allowing user to store and compute aggregations on Big data, there is still a lot of work to do in order to enable machine learning capabilities. In my thesis I would like to propose a suitable computational model, analyze its advantages over existing models and check its implementation on real-life data