机器学习中集成模型的应用问题研究

Research on application of integrated models in machine learning

文章综合运用了机器学习的各种理论对糖尿病的风险预测等问题进行了深度研究与分析,主要包括:(1)采用Logistic回归算法和决策树算法构建集成模型,最终发现集成模型能够在一定程度上解决单分类模型预测结果不稳定的问题,同时也能提升预测准确率。(2)风险预测模型由集成模型构建而来,综合比较预测准确率、精确度、召回率,发现集成模型的综合效果最好。

This paper comprehensively applied various theories of machine learning to conduct in-depth research and discussion on diabetes risk prediction and other issues. Its main work is as follows:(1) Logistic regression algorithm and decision tree algorithm are used to build an integrated model. Finally, it is found that the integrated model can solve the problem of unstable prediction results of single classification model to a certain extent and improve the accuracy of prediction.(2) The risk prediction model is constructed from the integration model, and the comprehensive comparison of prediction accuracy, accuracy and recall rate shows that the integrated effect of the integration model is the best.