A PARALLEL BIT-MAP BASED FRAMEWORK FOR CLASSIFICATION ALGORITHMS
Amila De Silva, Shehan Perera, Department of Computer Science & Engineering, University of Moratuwa, Katubedda, Sri Lanka
Bitmap representations have been abundantly used in data analytic queries for their ability to represent data concisely and for being able to simplify processing. For the same reasons, bitmaps are gaining popularity in Data Mining domain, with the arrival of GPUs, since Memory organisation and the design of a GPU demands for regular & simple structures. However, due to the nature of processing, use of bitmaps have largely been restricted to FIM based algorithms. We in this paper, present a framework based on bitmap techniques, which speeds up classification algorithms on GPUs. The proposed framework uses both CPU and GPU for the algorithm execution, where the core computing is delegated to GPU. We implement two classification algorithms Naive Bayes and Decision Trees, using the framework, both of which outperform CPU counterparts by several orders of magnitude
Data Mining, Classification, Bitmaps, Bit-Slices, GPU.