A novel HNN-DOC for automated agricultural ontology construction on climate factors

Abstract

Ontology has a huge potential for enhancing data association, executives, and understanding. One important criterion of domain ontologies is that they should accomplish a high level of inclusion of the domain concepts and concept relationships. However, the improvement of these ontologies is regularly a manual, time consuming, and frequently results in error. This problem leads towards the need for automation of ontology construction. Agriculture is a critical domain in our country faces problem due to the lack of knowledge on cropping pattern exclusively after the adverse effects of global warming. This paper gives a novel Hybrid Neural Network for Agricultural Ontology Construction (HNN-AGOC) for farmers through the Agro-Pedia dataset for predicting cropping automatically. The HNN-AGOC comprised of Convolutional Neural Network (CNN) for classifying the extracted features and Recursive Neural Network (RNN) for prediction in ontology construction. The algorithm was initially trained with the training dataset, and the performance analysis was performed on different performance metrics. The HNN-AGOC algorithm achieved the overall accuracy, precision, and recall values as 85.23%, 70.10%, and 80.24 % respectively.