Advanced gastric cancer survival prediction method based on ensemble learning

Abstract

The invention discloses a late gastric cancer survival prediction method based on integrated learning, and the method comprises the steps: obtaining case data, and carrying out the data screening, and obtaining a data set; constructing a prediction model and training the prediction model based on the data set to obtain a completely trained prediction model; and evaluating the training complete prediction model based on the confusion matrix and the working curve. Aiming at the characteristic of imbalance of survival samples of advanced gastric cancer, the clinical prognosis model established by adopting an integrated learning technology can solve the problem of overfitting. The advanced gastric cancer survival prediction method based on ensemble learning can be widely applied to the field of medical data processing.

Description

technical field [0001] The invention relates to the field of medical data processing, in particular to an integrated learning-based survival prediction method for advanced gastric cancer. Background technique [0002] Gastric cancer is one of the malignant tumors that seriously endanger human health, and its morbidity and mortality are among the highest in the world. Studies have shown that Helicobacter pylori (Hp) infection can lead to chronic gastritis, peptic ulcer and precancerous lesions, and Hp has become an independent factor affecting the high-risk groups of gastric cancer and a significant factor affecting the prognosis of gastric cancer. In addition, after a large number of clinical studies, the eradication of Hp can significantly reduce the incidence of gastric inflammation and even gastric cancer. At present, more than 50% of the residents in our country are positive for Hp infection, but at the same time, there is a lack of early screening and treatment of gast...

AI Technical Summary

Problems solved by technology

The traditional method is to use Cox proportional hazards regression to establish a nomogram for survival prediction. As a semi-parametric model, it is the most widely used multivariate analysis method in the field of medical survival analysis. However, Cox regression is mainly based on each characteristic variable and survival outcome. There is an assumption of linear correlation. Therefore, it oversimplifies the complex nonlinear relationship between feature variables, feature variables and survival outcomes. There is an over-fitting problem and the prediction accuracy is not high.

Images