Keyphrase Extraction for Technical Language Processing

Keyphrase extraction is an important facet of annotation tools that offer the provision metadata necessary for technical language processing (TLP). Because TLP imposes additional requirements on typical natural (NLP) methods, we examined keyphrase through lens a hypothetical toolkit which consists combination text features and classifiers suitable use in low-resource applications. We compared two approaches extraction: The first applied our toolkit-based methods used only distributional words phrases, second was Maui automatic topic indexer, well-known academic method. Performance measured against collections literature: 1153 articles from Journal Chemical Thermodynamics (JCT) curated by National Institute Standards Technology Research Center (TRC) 244 Task 5 Workshop Semantic Evaluation (SemEval). Both have author-provided keyphrases available; SemEval also reader-provided keyphrases. Our findings indicate approach competitive with when were removed text. For TRC-JCT articles, indexer reported F -measure 29.4 % while obtained 28.2 %. using Naïve Bayes classifier resulted 20.8 %, outperformed Maui’s 18.8