Preface
Computational approaches to subjectivity and sentiment analysis: Present and envisaged methods and applications

https://doi.org/10.1016/j.csl.2013.09.003Get rights and content

Abstract

Recent years have witnessed a surge of interest in computational methods for affect, ranging from opinion mining, to subjectivity detection, to sentiment and emotion analysis. This article presents a brief overview of the latest trends in the field and describes the manner in which the articles contained in the special issue contribute to the advancement of the area. Finally, we comment on the current challenges and envisaged developments of the subjectivity and sentiment analysis fields, as well as their application to other Natural Language Processing tasks and related domains.

Introduction

Recent years have witnessed a surge of interest in computational methods for affect, ranging from opinion mining, to subjectivity detection, to sentiment and emotion analysis. These methods typically focus on the identification of private states, such as opinions, emotions, sentiments, evaluations, beliefs, and speculations in natural language. While subjectivity classification labels text as either subjective or objective, sentiment classification adds an additional level of granularity by further classifying subjective text as either positive, negative or neutral, which is then further refined by emotion analysis by identifying the presence of emotions such as joy, anger, or fear.

In computational linguistics, the automatic detection of affect in texts is becoming increasingly important from an applicative point of view. Consider for example the tasks of opinion mining, market analysis, or natural language interfaces such as e-learning environments or educational/edutainment games. For instance, the following represent examples of applicative scenarios in which affective computing could make valuable and interesting contributions:

  • Sentiment analysis. Text categorization according to affective relevance, opinion exploration for market analysis, etc., are examples of applications of these techniques. While positive/negative valence annotation is an active area in sentiment analysis, a fine-grained emotion annotation could also contribute to the effectiveness of these applications.

  • Computer assisted creativity The automated generation of evaluative expressions with a bias on certain polarity orientation is a key component in automatic personalized advertisement and persuasive communication.

  • Verbal expressivity in human–computer interaction Future human–computer interaction is expected to emphasize naturalness and effectiveness, and hence the integration of models of possibly many human cognitive capabilities, including affective analysis and generation. For example, the expression of emotions by synthetic characters (e.g., embodied conversational agents) is now considered a key element for their believability. Affective words selection and understanding is crucial for realizing appropriate and expressive conversations.

The articles contained in this special issue are in their majority the extended versions of the best articles (as reviewed by the Program Committee) that have been presented at the 3rd Workshop on Computational Approaches to Subjectivity and Sentiment Analysis (WASSA 2012) – http://gplsi.dlsi.ua.es/congresos/wassa2012/. The event was organized in conjunction to the 50th Annual Meeting of the Association for Computational Linguistics, on July 12, 2012, in Jeju, Republic of Korea.

This edition has again shown that both the academic, as well as the industry communities have a great interest in the topics covered by the workshop. The large number of submissions and the tough review process ensured a high quality of the papers selected for presentation at the event. As such, the articles that are contained in this special issue regard important challenges in subjectivity and sentiment analysis areas, advancing the state of the art in these fields. They cover topics such as: multilingual sense subjectivity, multilingual sentiment analysis, lexicon and corpora building for subjectivity and sentiment analysis, emotion detection and contextuality and the applications of subjectivity and sentiment analysis to the detection of social and psychological phenomena. The types of texts on which the analysis is done vary from newspaper articles, to blogs and Social Media messages, thus exploring also the challenges posed by the structure of each of these text types.

In the following section, we present some of the most relevant work that has been recently conducted in subjectivity and sentiment analysis and describe the research trends in this field.

Section snippets

Recent trends in subjectivity and sentiment analysis

Subjectivity and sentiment analysis have been very active research topics in the past decade. Different authors have proposed methods to tackle the tasks from various types of texts and with diverse applications. Although much has been achieved in the field, some aspects still require additional tackling and further efforts are needed to expand the results to the multilingual and cross-lingual settings and new, informal types of texts. The articles contained in this special issue represent an

Special issue articles and their contribution to advancing research in subjectivity and sentiment analysis

In the following paragraphs, we briefly describe the ideas and contributions to the advancements of the state of the art brought by each of the articles included in the special issue:

  • The first article is entitled “Sense-level Subjectivity in a Multilingual Setting”, and is authored by Carmen Banea, Rada Mihalcea and Janyce Wiebe. This contribution deals with the issue of subjectivity classification in a multilingual setting. Particularly, the authors start from findings that subjectivity is a

Challenges and envisaged developments

Although much work has been carried out in the field of subjectivity and sentiment analysis, there are still many challenges to be overcome. First of all, as the papers in this special issue have shown, the approaches proposed in the field still require adaptation and much work to be done in order to be employed to other languages. Secondly, as new text types appear on the Social Web, the techniques to pre-process, as well as to tackle their informal style must be adapted, so as to obtain

Acknowledgments

First of all, we would like to thank the authors of the contributions submitted to WASSA 2012, and especially to those whose papers that were selected for this special issue, for their participation and contribution to the development of research in subjectivity and sentiment analysis. Further on, we would like to thank the WASSA 2012 Organizing Committee and Program Committee members, as well as their external reviewers, for their contribution to the workshop whose best contributions were the

References (22)

  • K. Ahmad et al.

    Multi-lingual sentiment analysis of financial news streams

  • B. Bader et al.

    Multilingual sentiment analysis using latent semantic indexing and machine learning

  • C. Banea et al.

    A bootstrapping method for building subjectivity lexicons for languages with scarce resources

  • C. Banea et al.

    Multilingual subjectivity: are more languages better?

  • C. Banea et al.

    Multilingual subjectivity analysis using machine translation

  • S. Deerwester et al.

    Indexing by latent semantic analysis

    Journal of the American Society for Information Science

    (1990)
  • A. Esuli et al.

    Enhancing opinion extraction by automatically annotated lexical resources

  • A. Go et al.

    Twitter sentiment classification using distant supervision

    Processing

    (2009)
  • T. Inui et al.

    Applying sentiment-oriented sentence filtering to multilingual review classification

  • L. Jiang et al.

    Target-dependent twitter sentiment classification

  • J. Kim et al.

    Evaluating multilanguage-comparability of subjectivity analysis systems

  • Cited by (67)

    • Improving sentiment analysis performance on morphologically rich languages: Language and domain independent approach

      2019, Computer Speech and Language
      Citation Excerpt :

      Previous studies have proposed several methods of addressing sentiment analysis in multiple languages. The most common approach is to adapt resources and tools developed for English and translate them into new languages utilizing bilingual dictionaries, machine translation, cross-lingual projections, or monolingual and multilingual bootstrapping (Balahur et al., 2014a). English sentiment lexicons were translated to other languages, i.e. Chinese (Ku et al., 2006), Romanian and Spanish (Banea et al., 2013), or Italian (Esuli and Sebastiani, 2011).

    • Tracking and recognizing emotions in short text messages from online chatting services

      2018, Information Processing and Management
      Citation Excerpt :

      Moreover, the semantics of words used within a chat session may be different from their common meaning. In addition to the above difficulties, it is notable that chatting is a type of continuous activity among the participants; consequently, an analysis should not be concerned solely with the textual and linguistic aspects (Balahur, Mihalcea, & Montoyo, 2014; Garas, Garcia, Skowron, & Schweitzer, 2012; Giachanou & Crestani, 2016). The emotion of a chatting message may relate to its context (e.g., location or previous messages) (Mohammad et al., 2016; Saif et al., 2016).

    • Distinguishing between facts and opinions for sentiment analysis: Survey and challenges

      2018, Information Fusion
      Citation Excerpt :

      Thus, very few task-specific corpora and dictionaries exist for subjectivity and sentiment analysis. In [26], the authors provide a review of subjectivity detection methods. They conclude that even Naïve Bayes trained on simple uni-grams can lead to good results.

    View all citing articles on Scopus
    View full text