A Study on Monitoring Water Pollution Events Based on Natural Language Processing and Social Media

With the rapid development of social and economic development, China’s environmental development is also extremely rapid, but there is still the phenomenon of sudden water pollution of the environment, which not only affects the cleanliness of the river waters, but also directly affects the residents of the normal life of water. Under the big data environment, a large amount of relevant information generated by news platforms or social media platforms when various sudden events occur can provide data support for event detection and information acquisition. In this paper, we use the named entity recognition technology in natural language processing to build a BERT-BiLSTM-CRF model for water pollution event monitoring using tweets from social media platforms as data sources. The experimental results show that the model extracts F1 values up to 0.9119, and the extraction results can provide additional data support for the discovery and monitoring of water pollution events around the world.