Metaheuristic Optimized Multi-Level Classification Learning System for Engineering Management

Multi-class classification is one of the major challenges in machine learning and an ongoing research issue. Classification algorithms are generally binary, but they must be extended to multi-class problems for real-world application. more complex than binary classification. In classification, only decision boundaries class known, whereas multiclass several involved. The objective this investigation propose a metaheuristic, optimized, multi-level system forecasting civil construction engineering. proposed integrates firefly algorithm (FA), metaheuristic intelligence, decomposition approaches, one-against-one (OAO) method, least squares support vector (LSSVM). enhanced FA automatically fine-tunes hyperparameters LSSVM construct optimized model. Ten benchmark functions used evaluate performance optimization algorithm. Two binary-class datasets related geotechnical engineering, concerning seismic bumps soil liquefaction, then clarify application problems. Further, uses cases engineering management verify effectiveness model diagnosis faults steel plates, quality water reservoir, determining urban land cover. results reveal that predicts plates with accuracy 91.085%, reservoir 93.650%, cover 87.274%. To demonstrate system, its predictive compared non-optimized baseline model, single (sequential minimal (SMO), Multiclass Classifier, Naïve Bayes, library (LibSVM) logistic regression) prior studies. analytical show promising project analytics software help makers solve applications.