Unsupervised Keyphrase Extraction with Multipartite Graphs
This addresses the problem of extracting keyphrases without supervision for text analysis applications, representing an incremental advance in graph-based methods.
The paper tackles unsupervised keyphrase extraction by proposing a model that uses a multipartite graph to encode topical information and leverage relationships between keyphrase candidates and topics, resulting in significant improvements over state-of-the-art graph-based models on three datasets.
We propose an unsupervised keyphrase extraction model that encodes topical information within a multipartite graph structure. Our model represents keyphrase candidates and topics in a single graph and exploits their mutually reinforcing relationship to improve candidate ranking. We further introduce a novel mechanism to incorporate keyphrase selection preferences into the model. Experiments conducted on three widely used datasets show significant improvements over state-of-the-art graph-based models.