چکیده مقاله
Accurate prediction of Alzheimer’s disease AD progression is a critical step towards timely intervention and personalized care Recurrent neural networks RNNs , particularly Long Short Term Memory LSTM models, have shown promise for modeling the longitudinal trajectories of clinical and cognitive variables However, they often struggle with the complexities of irregular time intervals and heterogeneous patient data In this paper, we propose an attention based deep learning framework that leverages Transformer like architectures to better capture relationships across multiple clinical visits, even when spaced unevenly in time By learning attention weights that emphasize the most informative visits and features, our model can more flexibly adapt to each patient’s unique progression pattern We validate our approach on a large, real world Alzheimer’s disease dataset, demonstrating superior performance over standard RNN models in predicting future dementia severity stages Moreover, the attention mechanism offers improved interpretability, enabling clinicians and researchers to identify which time points and clinical factors have the greatest impact on disease evolution Our results show that attention based methods represent a major step forward in understanding and forecasting the course of AD, ultimately supporting more efficient clinical decision making and patient management strategies
کلیدواژهها
نویسندگان
Hossein Choubin
Tarbiat Modares University, Faculty of Computer Science
شیوه ارجاع
Choubin, Hossein,1403,Attention-Based Deep Models for Irregular Alzheimer’s Disease Progression Prediction,The 5th international conference on artificial intelligence and its future prospects in electrical, computer, mechanical and telecommunication engineering sciences,Mashhad
ارائهشده در
مجموعه مقالات پنجمین کنفرانس بین المللی هوش مصنوعی و چشم انداز آینده آن در علوم مهندسی برق ، کامپیوتر ، مکانیک و مخابرات17 اسفند 1403 · مشهد