A Systematic Review on Sequential Pattern Mining-Types, Algorithms and Applications

Sequential Pattern Mining (SPM) is a vital area of data mining focused on uncovering meaningful patterns and subsequences within sequential data, such as time-series and transactional datasets. Despite its significance, existing reviews often overlook the rapid advancements and diverse applications of SPM techniques, leading to gaps in understanding their effectiveness and scalability. This paper provides a systematic review of SPM techniques and their applications from 2018 to 2024. A total of 1,440 articles were identified, with 31 selected based on rigorous screening criteria. The review categorizes these articles based on method type, dataset, metrics, and application domains, offering a structured analysis that emphasizes practical outcomes. Key findings reveal that while several SPM techniques have demonstrated significant improvements in accuracy and efficiency, challenges remain in scaling these methods to handle large datasets and complex patterns. The review highlights practical applications in fields such as special children assessment analysis, vehicle trajectory prediction in vehicular ad hoc networks (VANET), sitemap generation, and SPM based on uncertain databases. These insights underscore the need for robust algorithms tailored to address specific challenges within these domains. This work contributes a comprehensive overview of SPM, methodical classification of the reviewed literature, and identification of future research directions. By synthesizing current trends and practical implications, this study serves as a valuable resource for researchers and practitioners, guiding them in selecting appropriate methods for their specific needs and fostering further innovations in the field.