In this study, experiments were conducted to etch SiO₂ and Si₃N₄ by introducing N₂ at flow rates of 0, 2, 4, 6, and 8 sccm into a CF₄/O₂ plasma. OES (Optical Emission Spectroscopy) data were systematically collected and analyzed under each condition to understand the impact of N₂ addition on plasma chemistry. Machine learning techniques were applied to identify specific OES wavelengths that are critical to the etch rate and selectivity of both materials. Furthermore, the importance of the selected wavelengths was determined using XAI (Explainable Artificial Intelligence) methods.