Automated citation sentiment analysis using high order n-grams: a preliminary investigation

Abstract

Abstract:Scientific papers hold an association with previous research contributions (i.e. books, journals or conference papers, and web resources) in the form of citations. Citations are deemed as a link or relatedness of the previous work to the cited work. The nature of the cited material could be supportive (positive), contrastive (negative), or objective (neutral). Extraction of the author's sentiment towards the cited scientific articles is an emerging research discipline due to various linguistic differences between the citation sentences and other domains of sentiment analysis. In this paper, we propose a technique for the identification of the sentiment of the citing author towards the cited paper by extracting unigram, bigram, trigram, and pentagram adjective and adverb patterns from the citation text. After POS tagging of the citation text, we use the sentence parser for the extraction of linguistic features comprising adjectives, adverbs, and n-grams from the citation text. A sentiment score is then assigned to distinguish them as positive, negative, and neutral.In addition, the proposed technique is compared with manually classified citation text and 2 commercial tools, namely SEMANTRIA and THEYSAY, to determine their applicability to the citation corpus. These tools are based on different techniques for determining the sentiment orientation of the sentence. Analysis of the results shows that our proposed approach has achieved results comparable to the commercial counterparts with average precision, recall, and accuracy of 90%, 81.82%, and 85.91% respectively.

A decade of in-text citation analysis based on natural language processing and machine learning techniques: an overview of empirical studies

“…These methods are used to search for terms that reveal the sentiments and classify citation contexts into different polarity levels. The studies have utilised various feature-sets: the uni-gram feature seems to be very popular whereby higher values of n-grams lead to better classification results (Ikram, Afzal, and Butt, 2018). In the studies reviewed here, SVM and NB classifiers are the most frequently used models for sentiment classifications (Abu-Jbara et al, 2013;Athar, 2011).…”

“…In the dataset from the bioinformatics domain, 285 papers were randomly selected containing 3172 neutral, 702 positive, and 308 negative citations. Ikram and Afzal (2019) extracted different POS (nouns, proper nouns singular, proper nouns, determiners verbs, and adjectives) from citing sentences to Ikram et al (2018), they suggested applying a high value of n-grams (about n = 5) to achieve the best results.…”

“…As the annotated dataset compiled by Athar (2011) achieved promising accuracy, this dataset has been used by many studies as a gold-standard to analyse the polarity of citations. One of these studies has been published by Ikram, Afzal, and Butt (2018). They proposed a classification model and explored the usefulness of n-grams in citation-based sentiment analysis.…”

Machine Learning-Based Data Mining Method for Sentiment Analysis of the Sewol Ferry Disaster's Effect on Social Stress

“…The Semantria Application Programming Interface (API) is a cloud-based text analytics and sentiment analysis service based on advanced machine learning and natural language processing. It performs multilevel analyses of sentences incorporating parts of speech, assignment of a sentiment score from dictionaries, application of intensifiers, and determination of the final sentiment score based on machine learning techniques ( 22 ). Mostly used for business purposes, it is a tool that has recently seen increasing use in research for successful data mining ( 22 ).…”

“…It performs multilevel analyses of sentences incorporating parts of speech, assignment of a sentiment score from dictionaries, application of intensifiers, and determination of the final sentiment score based on machine learning techniques ( 22 ). Mostly used for business purposes, it is a tool that has recently seen increasing use in research for successful data mining ( 22 ). This tool provides an API for Excel and conducts its machine learning algorithm based on theme, entity, and category.…”