SailMutLoc: a sailfish-mutation enhanced optimization algorithm for student performance prediction

Feature selection enables educational data mining provide personalised support and enhance student performance. When using high-dimensional educational datasets, traditional feature selection techniques have premature convergence and inadequate feature sets. Inefficient feature selection occurs when these algorithms don’t balance exploration and exploiting. We propose the SailMutLoc algorithm, an integrated optimization method based on sailfish behaviour and enhanced by mutation operators and local search, to address these issues. SailMutLoc combines the Sailfish Algorithm (SFA)’s global search with the SCM’s fine-tuning precision and mutation-driven exploration. Iterative Local Search (ILS) increases local optimization results. SailMutLoc explores feature space without local optimum solutions since mutation makes things unpredictable. In studies using real-world educational datasets, SailMutLoc outperformed standard approaches in classification accuracy, computing time, and feature quality. In educational data mining, SailMutLoc can handle vast feature spaces and improve student performance forecasts.