Citation classification has been a research focus for several years, with the goal of identifying the author's citation intent by examining the text surrounding it. Due to the complexities involved in feature extraction, especially for multiple disciplines, most existing studies are confined to classifying citations within single domains. Besides,  the absence of standard benchmarks in the field has made comparing existing methods difficult. Additionally, the author's utilisation of implicit citation sentences when referencing a work introduces challenges in determining the optimal citation context boundary, thereby increasing the complexity of the process. Furthermore, the current methods heavily rely on the amount of annotated data used for training, making them data-centric. This presentation aims to explore automated ML and NLP methods that can be used to address the above shortcomings.