In the era of big data, social media platforms have become invaluable sources of information, offering vast amounts of data that can be analyzed to extract valuable insights. With the explosive growth of social media and the large volume of data it generates, noise has become a significant challenge in deriving meaningful insights. This article provides a comprehensive survey of noise elimination techniques in social media big data. In this paper, we review the literature on various noise elimination methods in social media big data and introduce a new classification of these strategies, such as preprocessing techniques, filtering methods, topic modelling, machine learning approaches, community detection, user-based filtering, outlier detection, anomaly detection, spam detection and related approaches. The survey examines the latest developments and recent advancements in noise reduction techniques, highlighting subcategories within each approach and their contributions to noise removal. Additionally, it discusses the challenges and opportunities associated with noise elimination in social media big data. This survey serves as a valuable resource for researchers and practitioners seeking an overview of noise reduction methods to improve the quality and reliability of social media big data analysis. Overall, this survey offers valuable insights into the landscape of noise elimination techniques in social media big data analysis, providing researchers and practitioners with a comprehensive understanding of existing methods and guiding future research in this important field.