Research and Application of a Question and Answer System for Bidding Based on Knowledge Graph

In order to solve the problem that the amount of bidding data is increasing and the relevant data is not easy to query, deep learning and other related technologies are used to solve the problem. Firstly, the content of the bid-winning announcement was crawled through crawler technology, and the obtained text data was cleaned and other processing, and the named entity recognition was carried out by constructing the BERT, BILSTM, and CRF fusion models. Experiments are carried out on the self-made bid-winning announcement dataset to prove that the proposed method can identify long sequence entities such as company names and improve the accuracy of entity recognition. Then, the Neo4j graph database was used for data storage and visualization to complete the construction of the graph. Finally, a Q&A system platform was built to achieve the effect of intelligent Q&A.