摘要
针对CLQUE算法聚类结果精确性不高的缺点,提出利用小波变换来生成自适应网格的方法对CLIQUE算法进行改进,将改进算法并行化以增强聚类维数升高时算法的可伸缩性,并将其应用于药品的销售预测。实验表明本算法聚类结果的精确性高,可伸缩性好,并且有效地降低了计算复杂度。
The classical CLIQUE algorithm is lost to a satisfying accuracy. So, a technique which generates adaptive grid by wavelet transform is put forward to improve it. The scalability of the improved algorithm is enhanced by par- allelization with the clustering dimensions increased. In the last, the algorithm is applied to drug sale estimation. The experimental demonstrate that the improved algorithm is more accurate, more scalable, and more efficient in decrease of computation complexity.
出处
《计算机科学》
CSCD
北大核心
2005年第3期216-218,共3页
Computer Science
基金
重庆市科技计划项目应用基础项目(7968)
