The Use of the BP Neural Network Model in Risk and Return Assessment of Investment Projects for Engineering Management

Abstract

This work aims to address the challenge of risk and return assessment in investment projects within the field of engineering management by proposing an innovative evaluation method based on a Back Propagation Neural Network (BPNN). By constructing and training the BPNN model, it successfully predicts the risk levels and returns of projects using a large dataset of historical engineering project data, significantly enhancing the accuracy of predictions and the scientific basis for decision-making. Experimental results show that, compared to traditional evaluation methods, the BPNN model demonstrates clear advantages in prediction accuracy, factor analysis, and decision support. The work also reveals limitations in areas such as data quality, model interpretability, and consideration of external factors, indicating directions for future research. The contribution of this work lies in showcasing the potential applications of artificial intelligence technology in engineering management. Moreover, this work provides new approaches and tools for risk assessment and return prediction of investment projects, with important theoretical and practical value.