- 學位論文
以FPGA實現DBSCAN分群演算法
The Implement of DBSCAN Clustering Algorithm Based on FPGA
摘要
在科技進步的時代，許多大型資料庫被拿來做研究，所以資料的分析與處理是相當重要的一環，這就使得分群演算法的選擇變得相當重要，進而如何縮短其分群速度也是一個值得思考的問題。 本論文中，我們提出一應用在數據資料分群上的一種具平行計算能力的DBSCAN(Density-based Spatial Clustering of Applications with Noise) 分群演算法的硬體架構。此架構可以尋找各群集密度較高的區域進行分群，並且能有效地去除雜訊，同時也可針對不規則的資料圖形進行分類。以硬體實現化減少分群處理及計算處理的資源複雜度並簡化其複雜度。且經由模擬軟體進行模擬測試。最後，經由模擬及FPGA實現驗證我們提出的硬體架構以驗證所提出的方法及架構是有效及可行的。
並列摘要
In the era of technological progress, many large databases to be used to do research, therefore, analysis and processing of information is a very important part, which makes the choice of clustering algorithm becomes very important, and then how to shorten its clustering speed is a question worth considering. In this thesis, we propose a hardware implementation with a parallel computing capability DBSCAN(Density-based Spatial Clustering of Applications with Noise) clustering algorithm on data clustering. The hardware implementation can search the high-density area for each cluster to classify the data, and remove the noise effectively. In addition, it can classify the irregular data graphics simultaneously. Furthermore, the hardware implementation can reduce the resources complexities substantially and compute the complexities of clustering processing and computing processing. Lastly, the examples are illustrated to show the validity and feasibility of the proposed clustering algorithm and hardware implementation by the simulation result and FPGA hardware architecture.