Lung Cancer Prognostic Model Evaluation with Continual Learning

5.5 KB dataset evaluates continual learning approaches for lung cancer prognostic models. The study proposes HCLMNet and employs methods like triplet-based contrastive learning and Cox Proportional Hazards models. It was authored by MD Ilias Bappi and last updated in March 2026.

Use Cases

• Compare Mean Absolute Error metrics between traditional deep learning and hybrid continual learning approaches like HCLMNet.
• Analyze the performance of Cox Proportional Hazards models in retaining prior knowledge from new patient data streams.
• Evaluate the effectiveness of Selective Sample Replay and Level Correlation Replay techniques for stabilizing learning in prognostic models.

Strengths

• Dataset is shared under a permissive CC BY 4.0 license.
• Focuses on a specific, high-impact application: lung cancer prognostic models.
• Includes evaluation of modern continual learning techniques like triplet-based contrastive learning.

Limitations

• Dataset is very small at 5.5 KB, indicating limited raw data or results.
• Specific data structure, column names, and row counts are unknown from the input.
• The description is highly technical and focused on methodology, with limited details on the underlying patient data.

Provenance

Source

figshare

Collection Method

Research output from an evaluation study of machine learning approaches.

Freshness

Last updated March 2026.