FPGA kernels for classification rule induction

摘要: Classification is one of the core tasks in machine learning data mining. One several models classification are rules, which use a set if-then rules to describe model. In this paper we present FPGA-based compute kernels for accelerating rule induction. The can be combined perform specific procedures induction process, such as evaluating coverage, or estimating out-of-bag-error. Since problems getting increasingly larger, there need faster implementations platforms that offer great potential mining FPGA (field programmable gate array), provides means implementing application accelerators.