重庆邮电大学学报（自然科学版）
重慶郵電大學學報（自然科學版）
중경유전대학학보（자연과학판）
JOURNAL OF CHONGQING UNIVERSITY OF POSTS AND TELECOMMUNICATIONS(NATURAL SCIENCE EDITION)
2010年
2期
242-247
,共6页
密度聚类%网格聚类%DBSCAN%聚类合并
密度聚類%網格聚類%DBSCAN%聚類閤併
밀도취류%망격취류%DBSCAN%취류합병
density-based clustering%grid-based clustering%density-based spatial chcstering of applications with noise ( DBSCAN )%clusters mergence
聚类已成为数据挖掘的主要方法之一,能够帮助人们在大量的数据中发现隐藏信息.目前最具典型的密度聚类算法是DBSCAN(density-based spatial clustering of applications with noise),它能够在空间数据库中很好地发现任意形状的簇并有效地处理噪声,但是它的计算复杂度相对较大.因此,采用划分数据集和聚簇合并方法,提出了一种基于密度和网格的高效聚类算法DGCA,并通过人工合成数据集和真实数据集对该聚类算法进行理论验证.实验结果表明该算法在效率性能和质量方面比DBSCAN都得到了提高.
聚類已成為數據挖掘的主要方法之一,能夠幫助人們在大量的數據中髮現隱藏信息.目前最具典型的密度聚類算法是DBSCAN(density-based spatial clustering of applications with noise),它能夠在空間數據庫中很好地髮現任意形狀的簇併有效地處理譟聲,但是它的計算複雜度相對較大.因此,採用劃分數據集和聚簇閤併方法,提齣瞭一種基于密度和網格的高效聚類算法DGCA,併通過人工閤成數據集和真實數據集對該聚類算法進行理論驗證.實驗結果錶明該算法在效率性能和質量方麵比DBSCAN都得到瞭提高.
취류이성위수거알굴적주요방법지일,능구방조인문재대량적수거중발현은장신식.목전최구전형적밀도취류산법시DBSCAN(density-based spatial clustering of applications with noise),타능구재공간수거고중흔호지발현임의형상적족병유효지처리조성,단시타적계산복잡도상대교대.인차,채용화분수거집화취족합병방법,제출료일충기우밀도화망격적고효취류산법DGCA,병통과인공합성수거집화진실수거집대해취류산법진행이론험증.실험결과표명해산법재효솔성능화질량방면비DBSCAN도득도료제고.
Clustering is one of the basic data mining tasks that can be used to help to understand the hidden information present in data sets density-based spatial clustering of applications with noise(DBSCAN), which is a typical density-based clustering algorithm, can detect arbitrary shaped clusters and handle noise well, but its computational complexity is unac-ceptable. In this paper, we present an efficient density and grid based clustering algorithm (DGCA) to enhance the per-formance of DBSCAN by partitioning data into clustering grids and merging clusters mutually. Synthetic data sets and SE-QUOIA 2000 benchmark are used for experimental evaluation to study the performance theoretically. Experimental results show that the efficiency and quality for clustering of the proposed algorithm are remarkably superior to those of DBSCAN.
