Publications by authors named "Alireza Manashty"
Electronic health records (EHR) are sparse, noisy, and private, with variable vital measurements and stay lengths. Deep learning models are the current state of the art in many machine learning domain; however, the EHR data is not a suitable training input for most of them. In this paper, we introduce RIMD, a novel deep learning model that consists of a decay mechanism, modular recurrent networks, and a custom loss function that learns minor classes.
View Article and Find Full Text PDFArticle Synopsis
- Predicting disease trajectories early helps doctors provide better treatment and avoid misdiagnosis, but it's tough due to issues like irregular patient data and long-term dependencies.
- The proposed solution, Clinical-GAN, uses a Transformer-based Generative Adversarial Network to forecast patient medical codes by treating them like sequences in language models, improving data interpretation with a multi-head attention mechanism.
- Evaluated on a large dataset from the MIMIC-IV database, Clinical-GAN outperforms existing prediction methods, showing its effectiveness in handling the challenges of medical data forecasting.