Fig. 2: Overview of the BOUND model development and validation process. | npj Digital Medicine

Fig. 2: Overview of the BOUND model development and validation process.

From: Interpretable machine learning model for digital lung cancer prescreening in Chinese populations with missing data

Fig. 2: Overview of the BOUND model development and validation process.

The BOUND model is composed of three key components: Bayesian network structure learning, parameter learning, and uncertainty inference. To make the model more user-friendly, a lung cancer risk scorecard was developed based on BOUND, enabling the assessment of high and low lung cancer risk. The model was trained on the training set and validated through both internal and external evaluations. The figure is created with BioRender.com.

Back to article page