This paper presents the results of a method for the visualization of the long-term prediction of research trending topics. Meaningful topics were identified among the words included in the titles of scientific articles. The title is the most important element of a scientific article and the main indication of the article’s subject and topic. We treated the titles’ words, which occur several times in articles cited in the analyzed collection, as the research trending topics. The longevity of the citation trend growth was the target for the machine learning algorithms. The CatBoost machine learning method, which is one of the best implementations of decision trees, was used. We conducted experiments on a scientific dataset that included 5 million publications from the top conferences in artificial intelligence and data mining areas to demonstrate the effectiveness of the proposed model. The accuracy rate of three-year forecasts for a number of experiments from 1997 to 2014 was about 60%. To visualize the forecast, the t-SNE and Word2Vec methods were used. Clusters of trending keywords on the semantic map helped to accurately identify promising directions. Two examples of forecast visualizations for the topic “Intelligent methods for data and image analysis” are presented. The presented visualizations serve as the analytical method for predicting topic trends and promising directions.

No abstract

No abstract

No abstract

No abstract