Ontology Synthesis Using Semi-Automatic Semantic AI Using Deep Learning and Reinforcement Learning

The need for automatic ontology modeling has become particularly significant in the era of Web 3.0, where ontologies serve as the structural metadata backbone. Despite their importance, certain domains such as environmental studies, though common—have been neglected and lack formal ontologies. To address this, the proposed model leverages auxiliary knowledge from Web 3.0, beginning with seed term pool generation from datasets and their classification using a Bi-LSTM model. Following this, the AdaBoost classifier is employed for dataset classification, while Pearson’s correlation coefficient aids in feature extraction through federation between two classifiers, enabling sedated learning. The integration of Google’s Knowledge Graph Search API further enhances the model by incorporating additional information, thereby strengthening its knowledge infrastructure. To ensure semantic accuracy, Morista’s Overlap Index is applied for semantics-oriented relevance computations, supporting reasoning through semantic similarity measures. Additionally, the Imperialist Competitive Algorithm is adopted as a metaheuristic optimization strategy. The use of YAGO along with Google’s Knowledge Graph API enriches the model with dense auxiliary knowledge, ultimately leading to high precision, and reliable estimation of F-measure with the lowest False Discovery Rate (FDR). In summary, the proposed framework named Ontology Synthesis using Semi-Automatic Semantic AI (OSSS) represents a state-of-the-art model for automatic ontology generation, specifically tailored for the domain of environmental journalism.