Review of the data-driven approach applications for analyzing thermophysical properties of composites

The paper presents an analysis of the possibilities and prospects of a data-driven approach for studying the thermo-physical properties of composite materials. An overview of key concepts and machine learning methods in materials science is provided, revealing their potential for studying and optimizing composite properties. The use of neural networks for predicting thermal characteristics is discussed. A comparison between the data-driven approach and traditional methods of analysis is made, identifying their advantages and disadvantages. Thermo-physical parameters of composites and modern methods of their measurement and prediction are described. The application of computer technologies for modeling the thermo-physical properties of composites and their manufacturing processes is considered. The research results may be useful for further development of materials science and composite materials production.