Journal

Volume 24, Issue 3 (September 30, 2023)

4 articles

• A Supervised Framework for Sentiment Analysis: A Two-Stage Approach
  by Prashantkumar M. Gavali & Suresh K. Shirgave
  J. CS. 2023, 24(3), 283-312;
  Abstract Sentiment analysis (SA) is a vital part of natural language processing (NLP). It involves analyzing textual data to determine expressed sentiment, be it positive or negative. Transformer-based models have gained popularity for sentiment prediction from the text. However, achieving promising results ... [Read more].
  Abstract Sentiment analysis (SA) is a vital part of natural language processing (NLP). It involves analyzing textual data to determine expressed sentiment, be it positive or negative. Transformer-based models have gained popularity for sentiment prediction from the text. However, achieving promising results with these models requires extensive training data and processing power. Additionally, when using pre-trained transformer models for sentiment analysis, they do not generate sentiment-specific text embeddings as they are trained on large general-purpose corpora. To address these challenges, this paper presents a two-step framework. Firstly, the framework aims to learn high-level sentiment-oriented embeddings of text. It generates embeddings with Siamese Network and compares them based on their sentiment class using the triplet loss function. This approach enables the framework to capture distinctions between positive and negative sentences, ensuring the generation of sentiment-specific embeddings. Secondly, it incorporates a classification layer on top of the embedding layer to enhance sentiment classification. Experimental results showcase the effectiveness of our proposed framework, surpassing baseline sentiment analysis results on various benchmark datasets. [Collapse]
• ECBFMBP: Design of an Ensemble deep learning Classifier with Bio-inspired Feature Selection for high-efficiency Multidomain Bug Prediction
  by Darshana Tambe & Lata Ragha
  J. CS. 2023, 24(3), 313-336;
  Abstract Prediction of software bugs from process logs, temporal access logs, behavior analysis, etc. requires estimation of a wide variety of high-density feature sets. Extracted feature sets must be able to classify these logs into different bug categories with high accuracy, and low complexity. To perform... [Read more].
  Abstract Prediction of software bugs from process logs, temporal access logs, behavior analysis, etc. requires estimation of a wide variety of high-density feature sets. Extracted feature sets must be able to classify these logs into different bug categories with high accuracy, and low complexity. To perform these tasks, a wide variety of Machine Learning Models (MLMs) are proposed by researchers, and each of them varies in terms of their performance-level nuances, functional advantages, contextual limitations, and application-specific future scopes. Upon analyzing these characteristics, it was observed that existing models are highly context-specific, and cannot be applied to multidomain bug analysis datasets. Moreover, existing models do not incorporate a dynamic feature selection method, which limits their accuracy performance under multiple bug classification applications. To overcome these issues, this paper proposes design of a novel Ensemble deep learning Classifier with Feature Selection for high-efficiency Multidomain Bug Prediction under different use cases. The proposed model improves bug representation performance by combining multiple feature extraction methods, including GWO-based novel feature selection techniques. The ensemble classification model, which combines Deep Random Forest, k Nearest Neighbor, Logistic Regression, Multilayer Perceptron, Support Vector Machine, and 1D Convolutional Neural Network classifiers, achieves higher accuracy, precision, recall, and low delay compared to existing models. The model also shows faster classification speeds than existing models and can be deployed for various real-time applications. This accuracy performance was compared with various state-of-the-art models, and it was observed that the proposed model showcases 4.5% higher accuracy, 3.2% better precision, 3.9% higher recall, and 5.5% faster classification performance, which was possible due to integration of intelligent feature selection process with high efficiency classification models under multidomain scenarios. [Collapse]
• Emotion Recognition in Comics: The Effect of Visual Morphemes in Visual Narrative Contexts
  by Hyorim Han & Jiyoun Choi
  J. CS. 2023, 24(3), 337-354;
  Abstract In comics, the term “visual morphemes” refers to one type of graphic structure that can be combined with other graphic elements to generate diverse meaning. For instance, the visual morphemes of whirlwind-shaped lines indicate confusion if they are placed above a person’s head. Prior empirical resea... [Read more].
  Abstract In comics, the term “visual morphemes” refers to one type of graphic structure that can be combined with other graphic elements to generate diverse meaning. For instance, the visual morphemes of whirlwind-shaped lines indicate confusion if they are placed above a person’s head. Prior empirical research has shown that such emotive visual morphemes do in fact help comic readers recognize the emotions of comic characters. However, there has been little evidence of the effect of emotive visual morphemes when emotion recognition is required in narrative contexts where multiple images are arranged to form a story, as opposed to when in solitary images of character-morpheme dyads. This study thus examined how emotive visual morphemes affect the identification of character’s emotions in narrative contexts consisting of three image panels. Results showed that emotion recognition was slower when the visual morphemes were not corresponding to the emotions of characters than when they were corresponding or when they were not provided at all. The findings thus add to our understanding of visual morpheme processing by providing empirical support for the emotive visual morpheme effects in the visual narrative structure. [Collapse]
• L2 Acquisition and Processing of Korean Direct Object and Oblique Relative Clauses by English Speakers
  by Hyomin Min & Seung-Ah Lee
  J. CS. 2023, 24(3), 355-400;
  Abstract This study examined the online comprehension of Korean direct object and oblique relative clauses (RCs) by English-speaking second language (L2) learners of Korean. A self-paced reading task and a follow-up picture selection task were conducted with adult English learners of Korean at intermediate a... [Read more].
  Abstract This study examined the online comprehension of Korean direct object and oblique relative clauses (RCs) by English-speaking second language (L2) learners of Korean. A self-paced reading task and a follow-up picture selection task were conducted with adult English learners of Korean at intermediate and low proficiency levels, together with a control group of adult native Korean speakers. The results from the picture selection task showed that oblique RCs were characterized by longer response times and lower accuracy rates than direct object RCs reflecting the noun phrase accessibility hierarchy effect. The outcomes of the self-paced reading task, however, revealed that in the two L2 learner groups, as well as the native speaker group, the total reading time was longer for direct object RCs than for oblique RCs (though the difference was statistically significant for learners with low proficiency only). The longer total reading time for the Korean direct object condition can largely be attributed to the momentary processing difficulty of the oblique argument marked by a postposition. This may be due to the possibility that ditransitive constructions are more difficult to process than transitive constructions. [Collapse]