Clustering and Thematic Analysis of News Content Using Machine-Learning Algorithms and Knowledge Graph

The amount of information on the Internet is increasing, and there is an increasing need for personalized content. Previously, only “quantity” was needed, but now higher quality and more accurate content are needed. This transformation requires media organizations to not only provide high-quality content, but also develop methods for content integration and dissemination. The news aggregation platform has an absolute advantage in the market and attracted a large number of advertisers with its continuous social media and precise content release. This article aims to analyze the connections and differences between news topics in the same article, and use machine-learning algorithms and knowledge graph to cluster and analyze news content, so as to identify the evolution process of topics corresponding to changes in citation motivation and help strengthen the recommendation effect of the article. The final experimental results demonstrated that the clustering of news content based on the method proposed in this paper and the clustering of thematic analysis had the best performance, and the accuracy of the algorithm was relatively high (when the data volume was 1.5 × 104 the mean square error of clustering and thematic analysis of news content based on this method was 1.701 × 10–4).