A New Semantic Relationship Extraction-based Ontology Construction for Agricultural Domain

Technological advancements have sparked diverse innovations in agricultural practices. One promising development is the construction of an agricultural ontology, which has the potential to propel the growth of both cross-domain agriculture and Natural Language Processing (NLP). However, such a domain-specific ontology requires a nuanced understanding of agricultural semantics and syntax. A standardized approach that leverages both syntactic and general semantic features for efficient ontology construction is required for data-scarce domains. The proposed Agriculture Domain-based Ontology construction (ADO) system takes the agricultural text documents as the input to construct an intricate ontology, weaving together domain based concepts and relationships. The given text undergoes anaphora resolution phase using the semantic-based method and domain terms are extracted using the linguistic filters. The relationships between the extracted terms are identified using semantic similarity and syntactic features. The various proposed text patterns in the research work are useful in extracting the relationships between agricultural entities in an effective way. The efficacy of the proposed method is compared with other prevailing systems and the results were encouraging. The proposed method shows a precision of $88.95 \%$ and a recall of $89.34 \%$.