Determining the semantic similarity of definitions by artificial intelligence for the needs of 3D Land Administration: a case of building units

Abstract

This article focuses on comparing selected terms connected to the building units across different standards and on testing the possibility of using the calculation of semantic similarity by artificial intelligence in the field of 3D Land Administration. In the first chapter, the reader is introduced to the issue itself. The second chapter is devoted to determining semantic similarity. First, it is mentioned that it is necessary to use NLP and the related text preprocessing steps. Subsequently, the methods of converting the text into numerical form (vectorization) and then the actual methods of calculation of the semantic similarity between two texts are described. In the third chapter, the reader will learn which standards and the corresponding terms and definitions have been selected. Subsequently, the results of the calculation of semantic similarity using artificial intelligences from OpenAI and Hugging Face are presented here. Finally, the results are compared with each other. In the last chapter, the entire text is summarized and some problems and possibilities for further research are discussed here.