RESUMO
Although diabetes mellitus is a complex and pervasive disease, most studies to date have focused on individual features, rather than considering the complexities of multivariate, multi-instance, and time-series data. In this study, we developed a novel diabetes prediction model that incorporates these complex data types. We applied advanced techniques of data imputation (bidirectional recurrent imputation for time series; BRITS) and feature selection (the least absolute shrinkage and selection operator; LASSO). Additionally, we utilized self-supervised algorithms and transfer learning to address the common issues with medical datasets, such as irregular data collection and sparsity. We also proposed a novel approach for discrete time-series data preprocessing, utilizing both shifting and rolling time windows and modifying time resolution. Our study evaluated the performance of a progressive self-transfer network for predicting diabetes, which demonstrated a significant improvement in metrics compared to non-progressive and single self-transfer prediction tasks, particularly in AUC, recall, and F1 score. These findings suggest that the proposed approach can mitigate accumulated errors and reflect temporal information, making it an effective tool for accurate diagnosis and disease management. In summary, our study highlights the importance of considering the complexities of multivariate, multi-instance, and time-series data in diabetes prediction.
Assuntos
Algoritmos , Diabetes Mellitus , Humanos , Fatores de Tempo , Diabetes Mellitus/diagnóstico , Aprendizagem , Aprendizado de MáquinaRESUMO
In the medical field, various clinical information has been accumulated to help clinicians provide personalized medicine and make better diagnoses. As chronic diseases share similar characteristics, it is possible to predict multiple chronic diseases using the accumulated data of each patient. Thus, we propose an intra-person multi-task learning framework that jointly predicts the status of correlated chronic diseases and improves the model performance. Because chronic diseases occur over a long period and are affected by various factors, we considered features related to each chronic disease and the temporal relationship of the time-series data for accurate prediction. The study was carried out in three stages: (1) data preprocessing and feature selection using bidirectional recurrent imputation for time series (BRITS) and the least absolute shrinkage and selection operator (LASSO); (2) a convolutional neural network and long short-term memory (CNN-LSTM) for single-task models; and (3) a novel intra-person multi-task learning CNN-LSTM framework developed to predict multiple chronic diseases simultaneously. Our multi-task learning method between correlated chronic diseases produced a more stable and accurate system than single-task models and other baseline recurrent networks. Furthermore, the proposed model was tested using different time steps to illustrate its flexibility and generalization across multiple time steps.