Power equipment triple construction method based on deep learning algorithm

The invention relates to the field of artificial intelligence. The invention discloses a power equipment triple construction method based on a deep learning algorithm. According to the electrical equipment triple construction method based on the deep learning algorithm, a more perfect electrical equipment knowledge graph triple extraction scheme is provided, namely, a mode layer is constructed in a top-down construction mode, and a data layer is constructed in a bottom-up mode under the guidance of the mode layer. According to the scheme, the characteristics of the power equipment text are clearly reflected, the design scheme of the power equipment triad can be perfected, and then the extraction efficiency and accuracy of the power equipment triad are improved; the invention further provides an entity relation extraction model comprising a bidirectional circulation network, an expansion gate convolutional neural network and a self-attention model. The model is constructed on the basis of a data layer construc