Fuzzy co-clustering can be improved if we handle two main problem first is outlier and second curse of dimensionality .outlier problem can be reduce by implementing page replacement algorithm like FIFO, LRU or priority algorithm in a set of frame of web pages efficiently through a search engine. The web page which has zero priority (outlier) can be represented in separate slot of frame. Whereas curse of dimensionality problem can be improved by implementing FCC_STF algorithm for web pages obtain by search engine that reduce the outlier problem first. The algorithm FCCM and FUZZY CO-DOK are compared with FCC_STF algorithm with merit and demerits on the bases of different fuzzifier used. FCC_STF algorithm in which fuzzifier fused into one entity who have shown high performance by experiment result of values (A1,B1,Vcj,A2,B2) seem to less sensitive to local maxima and obtain optimization search space in 2-D for web pages by plotting graph between J(fcc_stf) and Vcj.
翻译:如果我们首先处理两个主要问题,首先是超值问题,其次是维度的第二个诅咒问题,那么模糊的组合就可以改进。如果通过搜索引擎在一套网页框架中有效地实施FIFO、LRU或优先算法等网页替换算法,例如FIFO、LRU或优先算法,可以通过搜索引擎在一套网页框架中高效地执行一套网页框架框架框架框架框架框架。没有优先(超值)的网页可以在不同的框框中代表。虽然对维度问题的诅咒可以通过对通过搜索引擎获得的网页应用FCC_STF算法来改进,从而首先减少异常问题。 FCCM和FUZZY CO-DOK的算法与FCC_STF的算法进行了比较,在使用的不同发泡器的基础上将功绩和消减法进行了比较。FCC_STF,其中的模糊算法,在通过实验结果(A1,B1,Vcj,A2,B2)显示一个实体表现高性能,但对于本地最大值似乎不那么敏感,通过绘制J(fcc_stf)和Vcj之间的图,在2-D获得网页上的最优化搜索空间。