The invention relates to the technical field of
natural language processing, in particular to a place name entity recognition method based on
spatial relation perception, a medium and equipment, and the method comprises the following steps: constructing a corpus through
natural language sentences of place name entities in a network and pseudo sentences constructed by
processing a geographic
data set, and utilizing the pseudo sentences and the
natural language sentences, and further training a natural
language model by adopting comparative learning and
mask language modeling learning. According to the method, a Chinese
named entity recognition model is built based on a natural
language model, character-level feature representation is learned through word vector conversion, word-level feature representation is learned through the natural
language model, and new feature representation is generated by integrating the character-level feature representation and the word-level feature representation through a fusion embedding layer. And text context semantic features are extracted through a bidirectional long short-
term memory network, a
global optimal tag sequence is obtained through a
conditional random field, and place name entity recognition is realized. According to the method, the overall performance of the place name entity recognition model is improved.