BLSTM-CRF water conservancy and hydropower industry named entity recognition method fusing attention mechanism

• Main Authors: LI XIU'AN, YAN XINGTING, WANG KERAN, LI JIAYAO, GENG YONG, CAO WEIXING, FU ZHENYANG, BAO NANA
• Format: Patent
• Language: Chinese; English
• Published: 18.11.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention discloses a BLSTM-CRF water conservancy and hydropower industry named entity recognition method fused with an attention mechanism. The method comprises the steps of obtaining historical data, and performing text word segmentation to obtain an original text corpus; performing original sequence marking by using a BIO marking method to obtain preprocessed text data, and storing the preprocessed text data; after the text corpus is loaded, vectorizing the text by using a One-Hot one-hot coding representation method to obtain a vector expressing text semantics; carrying out feature extraction by utilizing a BLSTM layer of a constructed model and fusing an attention mechanism to obtain a semantic coding matrix; and calculating a state transition matrix of the semantic coding matrix by using the CRF layer of the constructed model, and correcting the text data by using the tag score to obtain final text data. According to the method, through the constructed entity recognition model, the named entities in