THE TECHNIQUE OF AUTOMATED IMAGE RESTORATION USING CONVOLUTIONAL NEURAL NETWORKS

The rapid growth of scientific knowledge and the ever-increasing volume of scientific publicationspose serious challenges to identify new trends and understand the changing research landscape. The formationof technological trends is necessary for the development and construction of developmentroadmaps at the national, sectoral and corporate levels. The task of identifying technological trends is animportant problem in the field of data analysis and machine learning. Well-known methods of analysis,including clustering by time factor, make it possible to form key phrases, but the task of forming trends,studying their characteristics and dynamics of their development does not currently have a satisfactorysolution. The solution to this problem involves: – creating a methodology for moving from key phrases todirectly naming new technological trends; – determination of the regularity of technology development ina given subject area; – determining the direction of future research. Solving these tasks will create aneffective decision support tool, reduce the time to identify a trend, assess the dynamics of its developmentand build roadmaps. In the presented work, a new approach to the formation of technological trends isproposed. The method is based on machine learning algorithms and natural language processing methodsand aims to overcome some of the limitations of traditional methods. In particular, the technique makes itpossible to identify complex relationships between various scientific concepts and provides a more accurateand comprehensive way to identify trends. The analysis of methods and methods for identifying trendsin scientific and technological development and their development based on keywords identified using amodel using time clustering is carried out. An algorithm for identifying trends is proposed