An Efficient Selection-Based kNN Architecture for Smart Embedded Hardware Accelerators

K-Nearest Neighbor (kNN) is an efficient algorithm used in many applications e.g. text categorization, data mining, and predictive analysis. Despite having a high computational complexity, kNN candidate for hardware acceleration since it parallelizable algorithm. This paper presents novel architecture implementation accelerator targeting modern System-on-Chips (SoCs). The adopts selection-based sorter dedicated that outperforms traditional sorters terms of resources, time latency, energy efficiency. has been designed using High-Level Synthesis (HLS) implemented on the Xilinx Zynqberry platform. Compared to similar state-of-the-art implementations, proposed provides speedups between 1.4× 875× with 41% 94% reductions consumption. To further enhance architecture, algorithmic-level Approximate Computing Techniques (ACTs) have applied. approximate accelerates classification process by 2.3× average reduced area size 56% real-time tactile processing case study. consumes 69% less accuracy loss than 3% when compared Exact kNN.