ABi-LT: Intelligent Medical Decision Model Clinical Decision Support System
Abstract
Accurate multi-label diagnosis and drug recommendation remain critical challenges in healthcare applications due to the complexity and scale of medical data. Traditional machine learning models often struggle to effectively capture long-term and global dependencies, leading to suboptimal performance and increased risk of overfitting. To address these limitations, this study proposes a novel hybrid architecture-Attention & Bi-LSTM & Transformer (ABi-LT)-which integrates Bi-directional Long Short-Term Memory (Bi-LSTM), attention mechanisms, and Transformer models. The Bi-LSTM component effectively captures sequential and contextual information, while the attention mechanism highlights salient features in the data by assigning adaptive weights. The Transformer component further enhances the model’s capability to learn global dependencies across large-scale datasets. Experimental results demonstrate that the ABi-LT model significantly outperforms traditional approaches in both multi-label classification accuracy and drug recommendation precision. This hybrid framework offers a scalable and robust solution for intelligent medical decision-making.