Evaluation and Prediction of Trends in the Development of Scientific Research Based on Bibliometric Analysis of Publications

The article proposes an approach to analyzing and predicting the thematic evolution of research by identifying an upward trend in keywords. Statistical analysis of the vocabulary of publications allows us to trace the depth of penetration of new ideas and methods, which can be set by the frequency of occurrence of words encoding whole concepts. The article presents a developed method for analyzing research trends and an article ranking algorithm based on the structure of a direct citation network. Data for the study was extracted from the Web of Science Core Collection, 6696 publications were collected for the experiment over the period 2005—2016 in the field of artificial intelligence. To evaluate the proposed method, 3211 publications were collected from 2017 to 2019. As a result, the method was evaluated by checking the presence of predicted keywords in the set of the most frequent terms for the period 2017—2019 and provided an accuracy of 73.33 %.