A Hash-based Hierarchical Algorithm for Massive Text Clustering

Text clustering is the process of segmenting a particular collection of texts into subgroups including content based similar ones. The purpose of text clustering is to meet human interests in information searching and understanding. This study proposes a new fast hierarchical text clustering algorithm HBSH (Hash-based Structure Hierarchical Clustering), which is suitable for massive text clustering. This algorithm uses hash table instead of numerical vectors as its input data. Compared with the other clustering algorithms, the HBSH performs the text clustering process without setting clustering center number and has minor space complexity in advance, which can achieve better performance. The experimental results illustrate that the average time of HBSH is faster than that of traditional text clustering algorithms.