Publications by authors named "Hedieh Mirzaieazar"
Background: Classification of binary data arises naturally in many clinical applications, such as patient risk stratification through ICD codes. One of the key practical challenges in data classification using machine learning is to avoid overfitting. Overfitting in supervised learning primarily occurs when a model learns random variations from noisy labels in training data rather than the underlying patterns.
View Article and Find Full Text PDFArticle Synopsis
- The study focuses on creating reliable mortality risk models to help doctors objectively assess patients' conditions in critical care settings, particularly within the ICU.
- It proposes a hybrid approach that combines clinical knowledge with advanced mathematical and machine learning techniques, using a tree-structured network for better interpretability of the model's predictions.
- The model is trained using graph-theoretic methods and shows effective validation across various hospital datasets, proving its ability to generalize well even when faced with different data structures.