Sentiment Analysis of Korean Using Effective Linguistic Features and Adjustment of Word Senses

2010 ◽  
Vol 14 (2) ◽  
pp. 33-46
Author(s):  
Hayeun Jang ◽  
Hyopil Shin
2017 ◽  
Vol 44 (2) ◽  
pp. 184-202 ◽  
Author(s):  
Adel Assiri ◽  
Ahmed Emam ◽  
Hmood Al-Dossari

Sentiment analysis (SA) techniques are applied to assess aspects of language that are used to express feelings, evaluations and opinions in areas such as customer sentiment extraction. Most studies have focused on SA techniques for widely used languages such as English, but less attention has been paid to Arabic, particularly the Saudi dialect. Most Arabic SA studies have built systems using supervised approaches that are domain dependent; hence, they achieve low performance when applied to a new domain different from the learning domain, and they require manually labelled training data, which are usually difficult to obtain. In this article, we propose a novel lexicon-based algorithm for Saudi dialect SA that features domain independence. We created an annotated Saudi dialect dataset and built a large-scale lexicon for the Saudi dialect. Then, we developed our weighted lexicon-based algorithm. The proposed algorithm mines the associations between polarity and non-polarity words for the dataset and then weights these words based on their associations. During algorithm development, we also proposed novel rules for handling some linguistic features such as negation and supplication. Several experiments were performed to evaluate the performance of the proposed algorithm.


Author(s):  
Elena G. Brunova ◽  
Yulia V. Bidulya ◽  
Alexander A. Gorbunov

The existing systems for accurate sentiment analysis are mainly based on statistical and mathematical principles. However, more promising are the works that are devoted to the study of the linguistic features of the evaluation expression. The results of this formalization can be applied both in the field of affective computing for further improvement of automatic systems and for linguistics and related sciences. The novelty of this study lies mainly in the development of an algorithm based on the identified linguistic rules. In addition, the research material is political discourse, which has not yet been studied enough by specialists of affective computing. The relevance of this work is justified by the growing need for categorization of information published on the Internet. The purpose of the study is to develop a system for machine sentiment analysis of English-language political texts, as well as to identify aspects and their distribution for subsequent use in enhancement. The article discusses the linguistic features of sentiment analysis and suggests a classification of linguistic units with sentiment potential in relation to levels of language structure. The results of an experiment on testing the operation of the sentiment analysis system, conducted on 300 news articles and user comments taken from reddit.com/r/politics, are also presented. The accuracy of the system is 92%. In addition, the selected 40 comments were manually marked up and tagged; during this process the expert identified 25 aspects. Furthermore, 3 formal patterns were identified in the distribution of aspect terms, which is necessary for creating an automatic system. The first peculiarity is that the aspect terms are repeated in two consecutive sentences. The second is that aspect terms are often the themes of sentences. Finally, the third — a high frequency of distribution of aspect terms at the beginning and end of the text (document) was revealed.


Author(s):  
María Pilar Salas-Zárate ◽  
Mario Andrés Paredes-Valverde ◽  
Miguel Ángel Rodríguez-García ◽  
Rafael Valencia-García ◽  
Giner Alor-Hernández

Author(s):  
Miguel V. Oliveira ◽  
Tiago de Melo

Identifying subjective sentences and classifying the polarity of subjective sentences are two important tasks in sentiment analysis. Besides being a hot topic, there is still a lack of resources to perform the mentioned sentiment analysis tasks in the Portuguese language, with its syntactic specificities. This paper describes the identified challenges and next steps in an initial study regarding the classification of subjectivity and polarity of sentences with a small set of syntactic features extracted directly from the text. Our approach reached satisfying results in experiments with two classic machine learning models in four datasets consisting of user reviews from different domains.


Mathematics ◽  
2021 ◽  
Vol 9 (12) ◽  
pp. 1449
Author(s):  
Tajana Ban Ban Kirigin ◽  
Sanda Bujačić Bujačić Babić ◽  
Benedikt Perak

This paper describes a graph method for labeling word senses and identifying lexical sentiment potential by integrating the corpus-based syntactic-semantic dependency graph layer, lexical semantic and sentiment dictionaries. The method, implemented as ConGraCNet application on different languages and corpora, projects a semantic function onto a particular syntactical dependency layer and constructs a seed lexeme graph with collocates of high conceptual similarity. The seed lexeme graph is clustered into subgraphs that reveal the polysemous semantic nature of a lexeme in a corpus. The construction of the WordNet hypernym graph provides a set of synset labels that generalize the senses for each lexical cluster. By integrating sentiment dictionaries, we introduce graph propagation methods for sentiment analysis. Original dictionary sentiment values are integrated into ConGraCNet lexical graph to compute sentiment values of node lexemes and lexical clusters, and identify the sentiment potential of lexemes with respect to a corpus. The method can be used to resolve sparseness of sentiment dictionaries and enrich the sentiment evaluation of lexical structures in sentiment dictionaries by revealing the relative sentiment potential of polysemous lexemes with respect to a specific corpus. The proposed approach has the potential to be used as a complementary method to other NLP resources and tasks, including word disambiguation, domain relatedness, sense structure, metaphoricity, as well as a cross- and intra-cultural discourse variations of prototypical conceptualization patterns and knowledge representations.


Author(s):  
Natalie Shapira ◽  
Gal Lazarus ◽  
Yoav Goldberg ◽  
Eva Gilboa-Schechtman ◽  
Rivka Tuval-Mashiach ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document