Association Link Network Based Core Events Discovery on the Web

Autor: Yang Liu, Hui Zhang, Xiangfeng Luo, Nazanin Borhan, Xiang He
Přispěvatelé: Chen, J, Cuzzocrea, A, Yang, L.T.
Rok vydání: 2013
Předmět:
Zdroj: CSE
Popis: As documents are explosively increasing in the era of big data, document clustering has been proven to be useful for organizing online document streams into events. However, extant studies on document clustering still suffer from the problems of high dimensionality, scalability and accuracy. In this paper, we will present a novel association link network (ALN) based document clustering method, which is an adaptive iteration splitting process to discover core events on the web. In the iteration, we first detect community structures from ALN; then, map documents to the associated community based on words relations in ALN; finally rebuild communities using the mapped documents. Compared to existing document clustering methods, the effectiveness of presented clustering method in automatically discovering the web events is proved by the experimental results on real data set.
Databáze: OpenAIRE