H.8. Document and Text Processing
Rojiar Pir Mohammadiani; Faezeh Hashemnia; Elham Moradizadeh; Soma Solaiman Zadeh
Abstract
Aspect-Based Sentiment Analysis (ABSA) has become a critical tool for extracting fine-grained insights from user opinions. This paper introduces DeGF-ABSA (DeBERTa-Gated Fusion for Aspect-Based Sentiment Analysis), a novel architecture that addresses key limitations in existing approaches by dynamically ...
Read More
Aspect-Based Sentiment Analysis (ABSA) has become a critical tool for extracting fine-grained insights from user opinions. This paper introduces DeGF-ABSA (DeBERTa-Gated Fusion for Aspect-Based Sentiment Analysis), a novel architecture that addresses key limitations in existing approaches by dynamically balancing global contextual features and aspect features. Unlike traditional methods that rigidly combine context and aspect representations—or transformer-based models lacking explicit mechanisms to disentangle aspect-specific signals—DeGF-ABSA leverages DeBERTa’s disentangled attention mechanism, which excels at modeling positional dependencies in technical texts, paired with a gated fusion layer. This layer adaptively weights the contributions of the context features that come from the [CLS] token, and the aspect-specific features come from the mean of aspect tokens. This helps in accurately determining the sentiment in complex sentences. Experiments on SemEval 2014 datasets achieve state-of-the-art results: 86.68% accuracy (84.50% F1) for Laptops and 91.43% accuracy (86.83% F1) for Restaurants.Cross-domain generalization is critical for aspect-based sentiment analysis, as domain-specific aspects and vocabulary vary significantly. Sentiment expressions also differ across domains, such as 'delicious' for food reviews versus 'fast performance' for electronics, requiring adaptable models to capture contextual nuances. Evaluating the DeGF-ABSA model's performance on datasets from domains beyond laptops and restaurants would provide valuable insights into its ability to generalize and its potential for broader applicability.
H.8. Document and Text Processing
Sh. Rafieian; A. Baraani dastjerdi
Abstract
With due respect to the authors’ rights, plagiarism detection, is one of the critical problems in the field of text-mining that many researchers are interested in. This issue is considered as a serious one in high academic institutions. There exist language-free tools which do not yield any reliable ...
Read More
With due respect to the authors’ rights, plagiarism detection, is one of the critical problems in the field of text-mining that many researchers are interested in. This issue is considered as a serious one in high academic institutions. There exist language-free tools which do not yield any reliable results since the special features of every language are ignored in them. Considering the paucity of works in the field of Persian language due to lack of reliable plagiarism checkers in Persian there is a need for a method to improve the accuracy of detecting plagiarized Persian phrases. Attempt is made in the article to present the PCP solution. This solution is a combinational method that in addition to meaning and stem of words, synonyms and pluralization is dealt with by applying the document tree representation based on manner fingerprinting the text in the 3-grams words. The obtained grams are eliminated from the text, hashed through the BKDR hash function, and stored as the fingerprint of a document in fingerprints of reference documents repository, for checking suspicious documents. The PCP proposed method here is evaluated by eight experiments on seven different sets, which include suspicions document and the reference document, from the Hamshahri newspaper website. The results indicate that accuracy of this proposed method in detection of similar texts in comparison with "Winnowing" localized method has 21.15 percent is improvement average. The accuracy of the PCP method in detecting the similarity in comparison with the language-free tool reveals 31.65 percent improvement average.
