Spiking Equilibrium Convolutional Neural Network for Spatial Urban Ontology

Urban analysis uses new data integration with computational methods to gain insight into urban methodologies. But the challenge is how to populate automatically from various urban documents. This paper proposes the ontology population problem by an ontology system for knowledge acquisition from textual resources and automatically populating the spatial urban ontology. Further, it aims to identify and extract useful textual terms and assign them to the predefined concepts (classes), instances (individuals), attributes (data properties) and relationships (object properties) of the urban ontology. The proposed ontology population system combines Natural language processing techniques and deep learning. Initially, the proposed work undergoes three major processes. They are the data acquisition, knowledge extraction, and ontology population processes. The deep learning model spiking equilibrium convolutional neural network (SECNN) is used to obtain high-quality information from text. The model’s performance is evaluated on precision, recall and F1-score metrics. The proposed SECNN attained a better precision value of 96.18%, recall value of 91.42% and F1-score value of 97.52%, respectively. Thus, the proposed SECNN model shows improved effectiveness over other models on spatial ontology.