Perioperative Immune-Inflammatory Analysis Method for Colorectal Cancer Based on Explainable Machine Learning

The single-controlled experimental method plays an important role in medical research, while it has a lot of limitations. With the rapid development of artificial intelligence, machine learning, and game theory are playing increasingly important roles. The research is about what factors infect the immune inflammatory response of colorectal cancer surgery. We propose an innovative model to predict immunity inflammatory response, and the model has more accuracy and more interpretability. Ensemble learning improves accuracy and reduces errors. The shap value explains which features are important and how the features explain the result. Many factors are considered, and the result shows that the anesthesia method has a significant effect. The time of the surgery has no significant effect on the immune inflammatory response, which means that the surgical schedule can be more flexible.