Paraphrase Detection using Semantic Relatedness based on Synset Shortest Path in WordNet

Lee, Jun Choi and Cheah, Yu-N (2016) Paraphrase Detection using Semantic Relatedness based on Synset Shortest Path in WordNet. In: International Conference on Advanced Informatics: Concepts, Theory and Applications, 16-17 August 2016, Parkroyal Penang Resort.

Preview

PDF Paraphrase Detection using Semantic Relatedness (abstract).pdf Download (119kB) | Preview

Abstract

This study presents a sentence-to-sentence semantic relatedness measures for paraphrase detection. The proposed measures adopt the shortest path between synsets in WordNet as the core to measure the relatedness between two sentences. The interlinked synsets in WordNet are based on the conceptual-semantic relation between two synsets. Thus the distance between two synsets in WordNet can be used to measure the semantic relatedness between two synsets. This study derived a sentence-level semantic relatedness using this feature to detect paraphrasing among sentences. The performance of the proposed semantic relatedness in paraphrasing is evaluated based on the accuracy and F-measures of the proposed measures in identifying paraphrase in Microsoft Research Paraphrase Corpus. The proposed method achieved 71.1% in accuracy and 81.8% in F-measures. The performance of the proposed method is compared with 6 paraphrase detection methods which include Salient Semantic Analysis and Second-order Co-occurrence Pointwise Mutual Information. In the comparison, the proposed method achieved the fourth highest accuracy and the second highest F-measure compare to other methods. This is a reasonable performance for the proposed semantic relatedness in paraphrase detection. Keywords— paraphrase detection, semantic relatedness, WordNet, synset shortest path.

Actions (For repository members only: login required)

View Item