The Role of Education in Building National Soft Power: An Empirical Analysis From a Global Perspective Using Deep Neural Networks

In recent years, deep learning technologies have rapidly transformed various sectors, particularly in fields such as healthcare, finance, and education. These advancements have led to significant achievements in predictive modeling, data analysis, and decision-making processes. However, there remains a need for more sophisticated models that can accurately capture complex relationships in dynamic, real-world datasets. To address this, this paper proposed a novel deep neural network (DNN) model designed to predict soft power based on educational factors. The proposed method involves collecting global educational data, including international student enrollment, research output, and the presence of top-ranked universities, and using this data to train a DNN model for soft power prediction. First, we preprocess the data by standardizing the features and handling missing values. Secondly, we build and train a DNN with multiple hidden layers to capture nonlinear relationships between educational factors and soft power scores. Thirdly, we evaluate the model's performance using several performance metrics, such as accuracy, RMSE, MAE, and precision. Finally, we compare the performance of our proposed DNN model with other machine learning algorithms, such as Random Forest and Support Vector Machines, demonstrating superior predictive accuracy. The experimental results indicate that the DNN model outperforms others, achieving the highest accuracy of 95.83%, exceptional sensitivity (94.45%), and specificity (96.26%). Its F1 Score of 95.56%, precision (95.87%), and recall (94.43%) reflect balanced and robust classification. The model also shows superior reliability with an MCC of 0.923 and an AUC of 0.978. Low error rates, including MAE (0.11), MSE (0.05), and RMSE (0.22), highlight its accuracy and predictive precision. The experimental results show that the DNN model significantly outperforms traditional machine learning models, providing a more accurate and robust prediction of soft power.