FPGA implementations of data mining algorithms

摘要: In recent decades there has been an exponential growth in quantity of collected data. Various data mining procedures have developed to extract information from such large amounts Handling ever increasing amount generates demand for computing power. There are several ways dealing with this demand, as multiprocessor systems, and use graphic processing units (GPU). Another way is field programmable gate array (FPGA) devices hardware accelerators. This paper gives a survey the application FPGAs accelerators mining. Three algorithms were selected survey: classification regression trees, support vector machines, k-means clustering. A literature review analysis FPGA implementations was conducted three algorithms. Conclusions on methods implementation, common problems limitations, means overcoming them drawn analysis.