Translation of Place Names Based on Knowledge Graph

Place names are important carriers of spatial information and attribute information for geographic entities. Transliteration of place names refers to the use of Chinese characters to translate place names in another language. However, current place name transliteration work suffers from issues such as time-consuming and inconsistent manual translation, as well as low accuracy in machine translation. Therefore, this article proposes a knowledge graph based method for English place name transliteration. This method mainly solves the core problem in place name translation: syllable optimization. By using a deep learning based phoneme generation method to generate place name proper phonemes, it is convenient to optimize the syllables of proper names using the constructed English place name knowledge graph. This knowledge graph-based syllable optimization effectively solves problems in machine translation, such as incomplete rule representation, infinite loops, one-to-many optimization results, weight setting, and nested rule optimization. Experimental comparisons conducted on Canadian place names demonstrate that the accuracy of the transliteration based on the knowledge graph can reach 91.2%, indicating that the proposed method improves the accuracy and efficiency of place name translation.