The need to leverage systematic mapping studies to develop and implement hierarchical evaluation models forces educators and AI developers to reevaluate algorithmic efficiency, computational feasibility, and contextual adaptability in order to identify what natural language processing methods are relevant and how they will be enacted in real-time tutoring environments and educational technology applications. Thisis where this research study aims to make a contribution, beyond introducing this comprehensive mapping framework, which presents empirical and theoretical articles dealing with natural language processing, conversation modeling, and interactional dynamicsof learning analytics, pedagogical dialogue analysis, deep learning architectures, or hybrid NLP methodologies. Following the systematic review of existing literature, a hierarchical classification of conversation analysis method development is compiled to link NLP advancements with these main terms, adaptive learning systems and intelligent tutoring frameworks. There are two main contributions of this systematic mapping study. One is that it is the first time to apply an Analytic Hierarchy Process (AHP) chart to show the ranking and prioritization of NLP techniques based on multiple evaluation criteria. Theother is that the framework embraces the idea of methodological synergy between supervised and unsupervised learning models, learner experience fulfillment and adaptive tutoring needs, and qualitative conversation analysis and quantitative performance evaluation that call for cross-disciplinary research in ITS development. This paper offers critical insights and reflections to provide academic researchers and industry practitioners with key information to best select and integrate NLP techniques for ITS optimization while being aware of computational constraints and scalability challenges.