iScience· 2025Q1
Chronic liver disease classification using deep learning with SHAP-optimized hybrid features
- 30citations
- Q1SCImago
- 2025year
Short summary
A novel deep neural network (DNN) framework achieved 92.50% accuracy in classifying liver disease by integrating SHAP-optimized hybrid features, outperforming traditional ML methods.
AI-generated from the title and abstract; the full text is not read.
Key points
- Developed a novel deep neural network (DNN) framework for liver disease classification.
- Achieved an average accuracy of 92.50% using 10-fold cross-validation.
- Integrated SHapley Additive exPlanations (SHAP) to identify influential features, enhancing model interpretability.
- The proposed DNN model outperformed traditional ML algorithms and state-of-the-art methods.
AI-generated from the title and abstract; the full text is not read.
Abstract
The liver is a vital organ responsible for essential functions, including digestion, metabolism, detoxification, and immunity. Liver disorders, whether due to disease, injury, or congenital conditions, pose serious health risks and require timely diagnosis for effective treatment and improved survival. Advances in machine learning (ML), particularly deep learning, have demonstrated significant potential for disease prediction, offering clinicians more accurate and efficient diagnostic tools. In this study, we propose a novel predictive framework based on a deep neural network (DNN) integrated with feature ranking and projection-based algorithms for accurate liver disease detection. To enhance model interpretability, SHapley Additive exPlanations (SHAP) were applied to identify the most influential features affecting predictions. Experimental results indicate that the proposed DNN model outperforms traditional ML algorithms and state-of-the-art methods, achieving an average accuracy (ACC) of 92.50% under 10-fold cross-validation. These results emphasis its potential to improve diagnostic ACC, support early intervention, and enhance patient outcomes.
The authors' abstract, as published at the source. iScience, 2025 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Health Information Management
Health Information ManagementHealth Professions