Graph-based Model for Topic Detection

In this paper, a novel graph-based model (GBM) is proposed for topic detecting. Different from existing statistical methods, our proposed model considers more semantic factors which combines named entity and dependency relation between words derived from a dependency parse tree. In our model, a graph is constructed for representing words and their association. By utilizing spectral clustering algorithm, we get clusters of words, each cluster represents a topic respectively. Our contribution includes as follows: modeling the topic detection problem as a graph-partitioning problem; proposing a new method of ranking the words association, and based on that, the document collection is represented as an undirected weighted graph. The performance of experiment task for dimensionality reduction and text classification indicates the feasibility and potentiality of our method.