Predicting COVID-19 patient recovery or mortality using deep neural decision tree and forest

Identifying patients at high risk of mortality is crucial for emergency physicians to allocate hospital resources effectively, particularly in regions with limited medical services. This need becomes even more pressing during global health crises that lead to significant morbidity and mortality. This study aimed to evaluate the effectiveness of deep neural decision forests and deep neural decision trees in predicting mortality among COVID-19 patients. To achieve this, we utilized patient data encompassing COVID-19 diagnosis, demographics, health indicators, and occupational risk factors to analyze disease severity and outcomes. The dataset was partitioned using a stratified sampling method. Nine machine learning and deep learning methods were employed to build predictive models.

Among the models, the deep neural decision forest outperformed others. Results indicated that using only clinical data yielded an accuracy of 80.7%, recall of 80.7%, precision of 75.7%, and F1-score of 74.8% by deep neural decision forest, demonstrating it as a reliable predictor of patient mortality. The model differs from other machine learning approaches for COVID-19 mortality prediction by combining the representational power of deep neural networks with the structured decision-making of decision forests, enhancing interpretability and performance using only clinical data without reliance on imaging or laboratory tests.