A Comparative Study on Representation Formalism Adoption for Semantic Knowledge Retrieval in Agriculture using Open Research Knowledge Graph

Abstract

Today’s focus on machine understandable designs that capture real world or domain information is required for solving complex problems in agriculture. Question answering in this domain require a robust semantic knowledge retrieval system that is vast in that domain and adequately represented knowledge for accurate and efficient semantic reasoning. This work is aimed at carrying out a critical review on the various representation methods, implementation approaches, and evaluation tools adopted in conducting agriculture-based researches. The reviewed articles are collected in the domain of agriculture. Open Research Knowledge Graph (ORKG) is adopted for creating comparisons used in this critical review including that of representation methods for multilingual machine translations. Visualizations from these comparisons are used in answering some competency questions surrounding this work and communicating the various results from this research. The results show 63 % analytical implementation work of most researches conducted in agriculture domain and 23 % automated. Knowledge graphs is mostly adopted in locations other than India and Nigeria. The report shows a high level of usage of general evaluation metrices such as accuracy, precision and recall for knowledge graph and ontology representations, pointing knowledge engineers to more researches on specific evaluation tools which are only being considered by very few of these researches. This will enhance the semantic knowledge retrieval procedures in agricultural domains as well as knowledge representation and reasoning for Semantic Web.