Ideological opinion clustering algorithm based on GTE text vector model with inverted index

The study of ideological public opinion on social media platforms has become a crucial focus in public administration and other domains as the internet develops rapidly. Conventional approaches of opinion analysis often suffer with low efficiency of large-scale data processing, inadequate semantic understanding of text, and noise interference. This work therefore suggests an ideological public opinion clustering method (GTE-ICA) grounded on GTE text vector model with inverted index. To enhance the clustering effect of ideological opinion data, the method combines the effective retrieval properties of inverted index with the strong textual semantic representation capacity of GTE model. Experimental data reveal that GTE-ICA has great noise robustness and performs very well on several conventional evaluation criteria. This work offers a reference guide for next development as well as a quick fix for handling ideological opinion data.