Abstract:
Author Profiling (AP) within the dynamic field of social media analysis is dedicated to predicting user attributes based on textual content. This systematic literature review presents a comprehensive analysis derived from 55 research papers, shedding light on the ever-evolving landscape of AP. Our meticulous investigation explores multifaceted aspects, including contextual features, ethical considerations, and the pivotal role of effective feature combinations. A significant focus of our study lies in unraveling the intricate relationship between linguistic features, machine learning algorithms, and their profound impact on the accuracy of age, gender and personality predictions. We delve into the challenges associated with dataset quality, cross-domain variations, and ethical concerns, systematically addressing these issues to provide a holistic view of the AP landscape. Our exploration extends beyond the conventional boundaries by discussing the relevance of sentiment analysis, natural language processing (NLP), and their alignment with social media analysis. We offer a nuanced perspective on the potential applications of AP across diverse languages and domains, emphasizing the adaptability and versatility of these profiling techniques. Throughout this review, the critical importance of effective feature combinations is highlighted, emphasizing their role in enhancing prediction accuracy. Contextual features, such as linguistic and stylistic elements, are examined in detail, providing valuable insights into their contribution to the evolving landscape of AP. The models showcase the ability to distinguish subtle differences in writing styles and language use, allowing for reliable predictions of author attributes. For instance, the models can discern age groups based on the frequency of identifying gender-age related language patterns with a high and achievable degree of accuracy. The review also underscores the ethical dimensions of AP, acknowledging the sensitivity of user data and the need for responsible practices in research and implementation. We tackle challenges related to biased datasets, privacy concerns, and potential misuse of profiling techniques, contributing to a more comprehensive understanding of the ethical considerations within the field. This comprehensive resource serves as an invaluable guide for researchers, practitioners, and policymakers, offering an up-to-date portrayal of the intricate and transformative landscape of AP. By synthesizing findings from diverse research papers, this review contributes to the academic discussion and provides practical insights for those engaged in the development and application of AP techniques.
Page(s):
1-1
DOI:
DOI not available
Published:
Journal: Second International Conference on Computing Technologies, Tools and Applications (ICTAPP-24), June 4-6,2024 (Abstract Book), Volume: 0, Issue: 0, Year: 2024
Keywords:
Machine learning algorithms
,
Social Media Analysis
,
Textual Data
,
Author Profiling
,
Age Prediction