Entity relationship extraction of operation and maintenance information based on deep Q-network algorithm

Due to the complexity of operational information entity objects, semi-structured states increase the probability of erroneous representation of operational information entity feature vectors, resulting in lower integrity of entity relationship extraction and even altering the original meaning of operational information. A method for extracting entity relationships of operation and maintenance information based on deep Q-network algorithm is proposed. Capture semantic associations and perform word segmentation on the feature vectors of operational information entities. Based on the deep Q-network algorithm, unified matching segmentation is performed to extract entity relationships in operation and maintenance information. The test results show that the Hits index of the extracted operational information entity relationships by this method is consistently above 0.85, indicating a stronger correlation between the extracted entities. The entity coding coverage index is maintained above 0.73, ensuring the integrity of information relationship extraction.