基于Logistic回归和Noisy-or模型的抑郁症风险预测研究

A Risk Prediction Model for Depression Based on Logistic Regression and Noisy-or

本文为了实现抑郁症的早期识别和检测,将Logistic回归模型与Noisy-or模型相结合,提出一种用于抑郁症早期识别和检测的风险预测模型LRANO。在该模型中Logistic回归用于计算不同风险因素对抑郁的概率贡献,Noisy-or模型将各种模型参数整合,形成最终的抑郁症风险预测模型。此外,通过在爱尔兰老龄化纵向研究数据库(TILDA)中进行验证,该模型的AUC值为0.7313,平均绝对误差为0.0887,表明了模型的有效性。

By combining the Logistic repression model with the Noisy-or model,a risk prediction model LRANO was proposed for the eprly identification and detection of depression.In Wis model,the Logistic repression model was taken to calculate the probabilistic contribution of different risk factors to depression,and the Noisy-or model was combined to integrate the various model parameters to form the final risk prediction model for depression.In addition,the AUC value of the model is 0.7313 and the average absolute error is 0.2887 through validation on The Irish Longitudinal Study on Aging(TILDA),which indicates the validity of the model.