This study explores the estimation of school performance using PCA and artificial intelligence. It attempts to integrate principal component analysis (PCA) and AI-based predictive models to determine the relative influence of different components on student performance. Traditional statistical methods are effective, but they often ignore the subtle interactions between different factors. These interactions greatly complicate problem solving and require more sophisticated methods. This methodology combines PCA with dimensionality reduction and machine learning techniques and analyses three interrelated components: (a) school grades, (b) positive effects and (c) negative effects. These components are aggregated to estimate performance using a weighted formula. Results show the effectiveness of this integrated approach, with linear regression outperforming neural networks. This study provides practical tools for educators while recognising the limitations of sample size and data completeness. Future research directions include integrating qualitative data and testing the methodology in different educational contexts.