The role of e-learning systems has become imperative in effectively educating masses of knowledge communities while maximizing the learner’s productivity. Barring this important role, e-learning systems face major challenges such as having context-aware and reusable learning contents. Furthermore, aspects oflearner profiling and categorization for deliverance of relevant learning contents, personalization and adaptive content recommendation to learners need to be focused. Currently, learning contents are static and not machine processable.Learner profiling may not fully comprehend the implicit as well as explicit characteristics of learners with subjective consideration of academic aspects at abstract level of granularity. Learner categorization techniques lack in dynamically considering the cognitive and inclinatory attributes of learners at finer level of granularity across the learning cycle. The learning contents offered may not accord with learning capacity of learners (lack personalization) with minimal support for content adaptivity.In proposed research, Ontology based Adaptive Semantic E-learning Framework (OASEF) is presented that exploits comprehensive set of learner attributes identified for effectively profiling the learners based on discriminative ones. Machine learning based dynamic and adaptive technique named Learner Categorization based on Hybrid Artificial Intelligence Techniques (LCHAIT) has been proposed for learner categorization. A supervised mode of learning was employed on a labeled data set modeled through a LearnerOntology. It has diverse learner’s profiles with implicit and explicit attributes pertinent to learner’s perspectives of demographics, academics, inclinations and behaviors. A comparative analysis of LCHAIT with three other machine learning techniques (Fuzzy Logic, Case Based Reasoning, and Artificial Neural Networks) is also presented. The learning contents maintained in the ontologies (CourseOntology, AssessmentOntoloy and DomainOntology) were recommended by considering the learner’s category to ensure personalization by a dynamic content recommender named Knowledge based Adaptive Semantic e-Learning Recommender (KASER). The efficacy of all categorization techniques was empirically measured while categorizing the learners based on their profiles through metrics of accuracy, precision, recall, f-measure and associated costs. These empirical quantifications assert LCHAIT as a better option than contemporary techniques as exhibited by greater accuracy of performance metrics. The performance of KASER was measured through degree of correctness in recommending the relevant learning contents compared with domain experts. Overall performance of OASEF was measured while recording the learner’s results spanning three years. The comparative analysis of proposed framework exhibits visibly improved results compared to prevalent approaches. These improvements are signified to the comprehensive attribute selection, learner profiling, dynamic techniques for learner categorization and effective content recommendation while ensuring personalization and adaptivity.