A Multiclass-based Classification Strategy for Rhetorical Sentence Categorization from Scientific Papers

Abstract

Abstract. Rapid identification of content structures in a scientific paper is of great importance particularly for those who actively engage in frontier research. This paper presents a multi terms of classification of r behind this approach is based on an observation that no single classifier is the best performer for classifyi our approach learns which classifiers for those categories and apply only the right classifier for classifying a given category. This paper obtain the category-classifier of the corresponding classifier using full This approach has been evaluated for identi categories on sentences collected from ACL experimental results show that the multi improve the classification performance over multi Keywords: acl-arc; classification strategies classification; rhetorical sentence categorization IntroductionKeeping abreast of the state researchers and reading new papers could be a daunting task proliferation of scientific publication. considered more effective is to provide readers with structured information that is extracted from a scientific paper. This structured information is represented as Rhetorical Document Profile (RDP) [1]. that readers want to know from a paper so that readers can identify the relevance of a paper just by reading its consisting of rhetorical slots. Each slot contains a collection of specific rhetorical category. Rhetorical sentence classification is the most important and major step in creating an RDP. This process is also known J. ICT Res. Appl., Vol. 7, No. 3, 2013, 235-249 Rapid identification of content structures in a scientific paper is of great importance particularly for those who actively engage in frontier research.a multi-classifier approach to identify such structures in terms of classification of rhetorical sentences in scientific papers. The idea approach is based on an observation that no single classifier is the best performer for classifying all rhetorical categories of sentences. Therefore, which classifiers are good at what categories, assign the classifiers for those categories and apply only the right classifier for classifying a This paper employs k-fold cross validation over training data to classifier mapping and then re-learn the classification model of the corresponding classifier using full training data on that particular category. This approach has been evaluated for identifying sixteen different rhetorical categories on sentences collected from ACL-ARC paper collection. The experimental results show that the multi-classifier approach can significantly improve the classification performance over multi-label classifiers.; classification strategies; multiclass approach; multi-label sentence categorization; scientific papers.Keeping abreast of the state-of-the-art of research topics is a must for researchers and reading new papers could be a daunting task with current fic publication. An alternative solution that can be considered more effective is to provide readers with structured information that is extracted from a scientific paper. This s...

Word Embedding for Rhetorical Sentence Categorization on Scientific Articles

“…For rhetorical sentence categorization, Liu [10] employed Word2Vec for seven rhetorical categories from Teufel [8] ('aim', 'textual', 'own', 'background', 'contrast', 'basis', and 'other'). In addition, Widyantoro, et al [3] implemented 16 rhetorical categories and features adopted from [8] and [11] in combination with a heterogeneous multi-classifier. Its average F-measure result was about 25%.…”

“…The abstract is the first section of a scientific article they read to find a summary of the information it contains [2]. It mostly contains a brief version of the study's objectives, methods, results and conclusion [3]. Nevertheless, readers cannot get all important information as needed only from the abstract because it does not reveal the correlation with other scientific articles.…”

“…Some works in rhetorical sentence categorization for scientific articles are argumentative zoning by classifying 7 rhetorical categories with Naïve Bayes [8], Maximum Entropy [9], Word2Vec [10], and 16 rhetorical categories adopted from [11] with a heterogeneous multi-classifier [3]. These works mostly adopted rhetorical sentence categorization from [8], which employs the meta-discourse feature.…”

“…However, word references in this meta-discourse are not always applicable for all scientific articles. There are some important words in the corpus of scientific articles [3] that are not included in the meta-discourse vocabulary from [8]. Some of these words are 'align', 'annotate', 'argument', 'aspect', 'concept', 'context', 'data', 'direct', 'document', 'domain', 'represent', and so on.…”

Automatic Title Generation in Scientific Articles for Authorship Assistance: A Summarization Approach

“…Previous researchers have done sentence classification as supervised learning task [12,[14][15][16][17]. In this study, of particular interest was rhetorical sentence classification where sentences are classified according to their information type/communicative purpose to judge their importance [12,18].…”