An Advanced E-Learning Recommendation System Utilizing Enhanced Wild Horse Optimization and Ensemble Classification

The growth of e-learning will result in the revolution in education. Here, a huge variety of materials and courses from any location in the globe can be accessed and it is facilitated by this revolution. The recommendations received by the learners is difficult to ensure that are both and engaging by using a lot of material. Data sparsity is the one of the issue that several current recommendation systems (RS) may face. This data sparsity occurs when limited data on interaction that will support for accurate predictions and over specialization when the recommendations are highly specialized. To deliver high-quality, personalized learning recommendations, advanced e-learning RS was introduced in the study, as it integrates the Enhanced Wild Horse Optimization (EWHO) with an ensemble classification technique. Fuzzy-C-Means Clustering (FCM) was initially employed by the system for the purpose of pre-processing, as it facilitates in detecting data dense regions and managing noisy data. To optimize system parameters as well as obtaining the best Fitness Values (FV), the EWHO is applied for enhancing the recommendation accuracy. Hybrid model of collaborative filtering (CF) and content-based (CB) approaches was integrated by RS. By using Jaccard Similarity Algorithm (JSA), user preferences based on item attributes and item similarities was analyzed by CBF. Then, Improved Genetic Algorithm (IGA) was employed, user preferences were predicted by CF, as it takes advantages of user interactions. An ensemble classification approach (ECA) is applied, which combines K-Nearest Neighbour (KNN), Kernel-based Support Vector Machine (SVM), and an Optimized Adaptive Neuro-Fuzzy Inference System (OANFIS) for the further enhancement of RS. Based on the needs of the individual learner, a personalized recommendation was facilitated by this integrated methodology. Robustness and accuracy was also ensured by this integrated methodology.