An Integrated Learning Framework for Hypertension Risk Prediction Using Machine Learning Algorithms

Abstract

In this study, a high-precision classification prediction framework fusing Random Forest algorithm and XGBoost algorithm is constructed. First, four single models, including logistic regression, Support Vector Machine (SVM), Decision Tree and Random Forest, were constructed and evaluated, and the evaluation results showed that logistic regression, SVM and Random Forest performed better in terms of interpretability and accuracy, while the Decision Tree was relatively poor; second, in order to break through the bottleneck of the performance of a single model, Random Forest was fused with XGBoost to construct a stacked integrated model, which adopts a two-layer architecture. Secondly, in order to break the performance bottleneck of a single model, Random Forest and XGBoost are fused together to construct a stacked integrated model, which adopts a two-layer architecture: the first layer trains the two base models independently and records the prediction probabilities, and the second layer splices these prediction probabilities with the original features to train the meta-model, which gives full play to the complementary advantages of the two features - Random Forest is based on the idea of Bagging, which can This architecture gives full play to the complementary advantages of the two features - Random Forest is based on the idea of Bagging, which can enhance the generalisation ability of the model and reduce the variance, and XGBoost is based on the idea of Boosting, which can effectively reduce the bias and deal with complex relationships. Through this stacking integration strategy, the models are deeply integrated to form a high-precision classification prediction system. Through multi-level model construction and integration optimisation, the framework significantly improves the accuracy, stability and generalisation ability of classification prediction, and shows excellent performance in classification prediction tasks, providing innovative prediction methods and technical paths for related fields.

Ähnliche Arbeiten