This thesis aims to enhance and deepen the understanding of security requirements by exploring how security requirements can be classified through a taxonomy. In order to do that, we have conducted a structured literature review on the database of IEEExplore to look for existing classifications and taxonomies of security requirements. We have found that there are many different existing classifications and taxonomies of security requirements. Among them are one-dimensional taxonomies and hierarchical taxonomies. While there are many similarities between the one-dimensional taxonomies, the hierarchical taxonomies differ greatly from each other. Additionally, we have constructed a classification of influence factors that affect the realization of security solutions of security requirements. Our classification framework was built utilizing the generative capabilities of the popular Large Language Model ChatGPT by generating hypothetical security scenarios and generating corresponding influence factors. Our classification consists of six different influence classes namely: Technical Factors, Organizational Factors, Threat Landscape Factors, Operational Factors, Regulatory Factors, and Human Factors.