The Evolution of Sentiment Analysis - A Review of Research Topics, Venues, and Top Cited Papers (1612.01556v4)

Abstract: Sentiment analysis is one of the fastest growing research areas in computer science, making it challenging to keep track of all the activities in the area. We present a computer-assisted literature review, where we utilize both text mining and qualitative coding, and analyze 6,996 papers from Scopus. We find that the roots of sentiment analysis are in the studies on public opinion analysis at the beginning of 20th century and in the text subjectivity analysis performed by the computational linguistics community in 1990's. However, the outbreak of computer-based sentiment analysis only occurred with the availability of subjective texts on the Web. Consequently, 99% of the papers have been published after 2004. Sentiment analysis papers are scattered to multiple publication venues, and the combined number of papers in the top-15 venues only represent ca. 30% of the papers in total. We present the top-20 cited papers from Google Scholar and Scopus and a taxonomy of research topics. In recent years, sentiment analysis has shifted from analyzing online product reviews to social media texts from Twitter and Facebook. Many topics beyond product reviews like stock markets, elections, disasters, medicine, software engineering and cyberbullying extend the utilization of sentiment analysis

• The paper presents a comprehensive bibliometric review using text mining and LDA to classify 6,996 sentiment analysis papers and trace its evolution.
• The study identifies a shift from product reviews to diverse applications including social media, finance, and public health.
• The research highlights the dispersion of publication venues and the substantial impact of top-cited works on shaping the field.

Overview of "The Evolution of Sentiment Analysis - A Review of Research Topics, Venues, and Top Cited Papers"

This paper offers a comprehensive bibliometric and literature paper on sentiment analysis, a rapidly expanding field within computer science. Utilizing a substantial corpus of 6,996 papers from Scopus, the paper employed a blend of text mining and qualitative analysis techniques to map out the development and current state of sentiment analysis research. The review addresses historical roots, publication venues, research topics, and influential papers, providing an in-depth perspective suitable for experienced researchers in the field.

Historical Context and Publications

The roots of sentiment analysis trace back to early 20th-century studies on public opinion and linguistic subjectivity in the 1990s. However, the field only began to flourish in the mid-2000s with the surge of subjective online content. Notably, 99% of sentiment analysis papers have been published post-2004, underscoring its rapid growth. A significant challenge for researchers is the dispersion of work across numerous venues, with the top-15 publication venues accounting for only about 30% of the total output.

Key Findings in the Literature

The authors present a taxonomy of sentiment analysis research topics, highlighting a shift from product reviews to a broader array of applications, including social media, finance, and public health. The taxonomy was developed using Latent Dirichlet Allocation (LDA) for text clustering, further refined through qualitative coding. This comprehensive classification underscores the diverse applicability of sentiment analysis across various sectors.

Research Topics and Methodologies

The paper categorizes research into areas related to data sources, analytical methods (e.g., machine learning, natural language processing), and application domains. A word cloud analysis further reveals evolving trends, such as the pivot from online reviews to social media content (e.g., Twitter and Facebook) in recent years.

Citation Analysis

The impact of sentiment analysis research is evident in its citation patterns. While over half of the papers remain uncited, the most influential works have a substantial citation footprint, exceeding those of established fields like software engineering. The analysis highlights pioneering works and reviews that have shaped the field, categorized into five main groups.

Implications and Future Research

This paper highlights sentiment analysis's expansive role in understanding human opinion across various contexts. It indicates a likely future where sentiment analysis integrates more naturally with broader affective computing methodologies, potentially leveraging multi-modal data sources like audiovisual content. The field's sustained growth may hinge on developing more sophisticated analytical tools and techniques.

Conclusion

The paper offers valuable insights into sentiment analysis, mapping its developmental trajectory and providing a detailed classification of its research landscape. It serves as a vital resource for scholars seeking an informed overview of the field, elucidating both the historical context and current research dynamics. While sentiment analysis continues to evolve, its broadening scope signifies its central importance in computational approaches to understanding human emotion and opinion.