Abstract Textual data are among the most popular data types that provide useful knowledge today. The need to extract insights from text has initiated and overseen the progression of research in the text mining domain. The challenge of addressing textual data comes from the complexity of text processing. The nature of text presents challenges that demand sophisticated processing methods. Subsequently, the degree of complexity involved in this field drives researchers to continuously improve the existing feature selection and extraction techniques, highlighting the essential work that is necessary to propel the field forward. Unlike the existing reviews that focus on individual components such as preprocessing, feature selection, or feature extraction in isolation, this study offers a unified perspective by integrating all three aspects. This review explores the feature selection and extraction methods advances achieved in text mining over the last decade. The focus of this investigation is the development of novel approaches that have been designed to address the complexities that accompany textual data. The comprehensive examination includes identifying the current trends and significant breakthroughs that characterize this dynamic field, thus providing a comprehensive view of the ongoing efforts to refine and optimize these methodologies. The presented explanation of the details associated with feature selection and extraction not only keeps researchers up to date on the latest developments but also contributes significantly to the ongoing refinement and growth exhibited by text mining techniques.
Ayash et al. (2025) studied this question.