Extraction Algorithm of Opinion Information Based on Double Embedding CMLA

With the popularity and development of the Internet, people have recorded a large number of comment texts on government service hotlines, social media and other network platforms to express their opinions. For the task of viewpoint information extraction, this paper proposes a joint viewpoint information extraction algorithm based on dual-embedding and coupled multi-layer attention mechanism. Combining dual-embedding word representation and coupled multi-layer attention mechanism improves the extraction of aspect words and opinion words, and achieves joint extraction of their correspondence.