Helicobacter pylori infection prediction method and system based on machine learning

The invention discloses a helicobacter pylori infection prediction method and system based on machine learning, and the method employs binary Logistic regression to carry out factor screening, further employs a decision tree to explore the interaction between factors, adds the interaction items obtained by the decision tree to generalized linear regression, obtains a prediction model, and carries out the prediction of helicobacter pylori infection. Compared with a conventional linear prediction model, the AUC of the final result of the prediction model is improved, it is proved that the prediction model can more effectively predict whether a patient is infected with the helicobacter pylori or not, and therefore the efficient auxiliary effect is achieved for prevention and treatment work of the helicobacter pylori. 本发明公开了一种基于机器学习的幽门螺旋杆菌感染预测方法及系统，本方法使用二元Logistic回归进行因素筛选，并进一步采用决策树对因素间交互作用进行探索，将决策树获取的交互项添加至广义线性回归，得到预测模型，相较于常规的线性预测模型，本预测模型最终结果的AUC有所提升，证明本预测模型能够对患者是否感染幽门螺旋杆菌进行更为有效的预测，从而为幽门螺旋杆菌的防治工作起到高效辅助作用。