Study on the Construction of Chronic Disease Risk Assessment and Early Warning Model by Integrating Multi-Source Physical Examination Data

In today’s era of rapid development of computer technology, artificial intelligence and intelligent algorithms are increasingly used in various fields. The study aims to construct and evaluate a chronic disease risk assessment and early warning model that integrates multi-source physical examination data. Aiming at the limitations of traditional methods in dealing with nonlinear relationships and high-dimensional data, the study adopts the convolutional neural network algorithm to construct an assessment model that can integrate physiological indexes, blood indexes, and lifestyle data through the steps of data preprocessing, feature selection, and multi-source data fusion. The outcomes of the study indicate that the classification accuracy of the proposed model is better than the differential evolutionary backpropagation neural network algorithm in the task of chronic disease classification, especially on disease types 1, 3 and 5, where the performance is more significant. Moreover, the risk assessment value for patient $\mathbf{P 0 0 5}$ was $\mathbf{0. 9 5}$, while that for patient $\mathbf{P 0 0 8}$ was $\mathbf{0. 5 2}$, showing the effectiveness of the model in quantifying the risk of chronic diseases in patients. The studied model provides a more accurate and reliable tool for chronic disease risk assessment, which has important clinical application value and research significance.