期刊名称：International Journal of Computer Science and Network
印刷版ISSN：2277-5420
出版年度：2013
卷号：2
期号：6
页码：179-183
出版社：IJCSN publisher
摘要：Discovery the suitable quantity of huddle to whichcredentials should be separation is vital in manuscript huddle.In this dissertation, we suggest a fresh approach, namelyDPMAP(Dirichilet Process Model Attribute Partition), torealize the embryonic huddle construction based on the DPMmodel lack in require the amount of huddle as key. Elementsclassify into two classes, important expressions and unmatchedterms. To infer document album constitution and separationdocument words at the equivalent time by using Variationassumption algorithm. The assessment sandwiched betweenour scheme and modern manuscript huddle method explainsthat our method is powerful and helpful for manuscript huddle
