NiO films with different thicknesses (100, 150, 200, 250, 300 and 400nm) were grown on glass substrates using the RF Magnetron sputtering method and their optical transmittance properties were analysed with a spectrophotometer. An innovative aspect of this work was the application of machine learning techniques used to derive new insights from experimental data. Four different machine learning algorithms -ANFIS, Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Gaussian Process Regression (GPR)- were tested. While the models were trained using films of different thicknesses, a randomly selected 75% of the whole dataset was used for model testing and the remaining 25% of the films were used for testing the models. Among these, ANN and GPR models were found to be the most successful models. Using these models, the energy band gaps were estimated at 1nm intervals and the values ranged from approximately 3.50eV to 3.76eV.
Read full abstract