1. Sandi Božičković, Saobraćajna i elektro škola, Doboj, Republika Srpska, Bosna i Hercegovina,
Republic of Srpska, Bosnia and Herzegovina
2. Ratko Garić, Fakultet za profesionalne studije za menadžment i poslovne komunikacije, Beograd, Serbia
The electrical properties of solid materials represent a set of physical characteristics that describe the behavior of a solid material in the presence of an electric field or electric current. They are very important for the selection of materials in electrical engineering, electronics, energy, mechanical engineering and modern technical systems. The most important electrical properties are: electrical conductivity (σ), electrical resistance (R), specific electrical resistance (ρ), dielectric properties, dielectric strength, electrical polarization, dielectric losses, etc.
This paper investigates the application of artificial intelligence to predict the electrical conductivity, electrical resistivity, band gap, and dielectric constant of solid materials. Machine learning models are developed using data on the chemical composition, crystal structure, temperature, and other physical characteristics of the materials. The performance of regression model, random forest (RF), and artificial neural network (ANN) models is evaluated using the coefficient of determination (R²), mean absolute error (MAE), and root mean squared error (RMSE) metrics. The proposed approach enables the identification of complex relationships between material structure and electrical properties, accelerates material characterization, and facilitates the selection of functional materials for applications in sensors, power electronics, and IoT (Internet of Things) devices.
Ključne reči :
Tematska oblast:
Datum:
24.08.2026.
Contemporary Materials 2026 - Savremeni Materijali
Datoteka uz rad
Datoteka uz rad