International Journal of Clinical Case Reports, 2025, Vol.15, No.5, 209-218 http://medscipublisher.com/index.php/ijccr 209 Research Insight Open Access Comparative Study on Construction Methods of Chronic Disease Prediction Models Based on Big Data Jingqiang Wang Institute of Life Science, Jiyang College of Zhejiang A&F University, Zhuji, 311800, Zhejiang, China Corresponding email: jingqiang.wang@jicat.org International Journal of Clinical Case Reports 2025, Vol.15, No.5 doi: 10.5376/ijccr.2025.15.0022 Received: 13 Jul., 2025 Accepted: 18 Aug., 2025 Published: 23 Sep., 2025 Copyright © 2025 Wang, This is an open access article published under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Preferred citation for this article: Wang J.Q., 2025, Comparative study on construction methods of chronic disease prediction models based on big data, International Journal of Clinical Case Reports, 15(5): 209-218 (doi: 10.5376/ijccr.2025.15.0022) Abstract This study explores various modeling methods for chronic disease prediction using big data, with a focus on chronic disease prediction. The basic theories of prediction and commonly used data types are introduced, including clinical indicators, genetic information and lifestyle-related data. The application situations of statistical models, machine learning models and deep learning models were compared. Statistical models are easy to understand, but they still have shortcomings when dealing with high-dimensional and nonlinear data. Machine learning models perform well in identifying complex patterns and integrating different data, but they rely on feature selection and parameter adjustment. Deep learning models have advantages in handling multimodal data and time series prediction, but they require more data and computing resources. This study also mentioned the evaluation criteria of the model, issues related to data quality and privacy protection, as well as the challenges in terms of acceptance and interpretability in clinical practical applications. Overall, the analysis results show that different models each have their own advantages. Future research should focus on the application of hybrid modeling, multi-source data fusion, reinforcement learning and causal reasoning in clinical practice. Keywords Big data; Chronic disease prediction; Machine learning; Deep learning; Model comparison 1 Introduction Chronic diseases such as heart disease, diabetes, cancer and kidney disease are important causes of global morbidity and mortality. The large number of patients, long treatment cycle and high cost have exerted great pressure on public health. With the aging of the population and changes in lifestyle, the problem of chronic diseases has become more serious. Therefore, the earlier the detection and intervention, the better the therapeutic effect can be improved and the cost can be reduced (Rashid et al., 2022). Traditional diagnostic methods often struggle to cope with the complexity of chronic diseases and the coexistence of multiple diseases. Therefore, more advanced predictive approaches are needed (Kim et al., 2023). In the medical field, the use of big data-including electronic medical records, management information, wearable device data and large-scale clinical data-is changing the way diseases are predicted and decisions are made. Integrating and analyzing health data from different sources can establish predictive models, identify high-risk groups, and support early diagnosis and personalized treatment. With the help of machine learning and artificial intelligence (AI) technologies of big data, complex patterns can be better discovered, the accuracy of prediction can be improved, and active health management can be promoted (Tsai et al., 2025). However, data quality, the interpretability of results, and the integration of multi-source data remain issues that need to be addressed (Liu et al., 2023). This study will explore the methods of establishing chronic disease prediction models using big data, with a focus on analyzing the advantages and disadvantages of machine learning, deep learning, and hybrid methods. The innovation points include designing a multi-task and multi-modal learning framework capable of simultaneously predicting multiple diseases, adopting more advanced feature engineering and integration strategies, and combining structured and unstructured data to enhance the prediction effect. Through experiments on different datasets and disease types, this study aims to provide practical references for researchers and clinical personnel, helping them select and apply more appropriate modeling methods in actual medical practice.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==