Article: Comparative evaluation of machine learning-based algorithms for clinical treatment prediction using EHR data from China Journal: International Journal of Internet Protocol Technology (IJIPT) 2026 Vol.19 No.3 pp.118 - 129 Abstract: The study aimed to assess a variety of Machine Learning tools for predicting treatment outcomes for patients based on the patient's electronic health record. The research framework is about supervised learning and ensemble methods, where algorithms classify the status of patient care learning from clinical data with their labels. Eight ML models were used for comparison, which include Decision Tree, Random Forest, Gradient Boosting, SVM, Gaussian Naive Bayes and Convolutional Neural Networks. Each model had been optimised by specific hyperparameter tuning, and their performance was assessed using metrics like mean absolute error, mean squared error, Matthew's correlation coefficient, F1-score and Area under the Curve of the Receiver Operating Characteristic. The evaluation metrics indicated that the best models are CNN, and AdaBoost. The comparative analysis serves as a reference for identifying apt ML methodologies in clinical decision support and opens a new path towards devising accurate and personalised treatment planning. Inderscience Publishers - linking academia, business and industry through research

Title: Comparative evaluation of machine learning-based algorithms for clinical treatment prediction using EHR data from China

Authors: Gui Wu; Shi Qian; Xin Luo; Rong Bie; Dandan Qu

Addresses: School of Health Services and Management, Hubei Preschool Teachers College, Wuhan, Hubei Province, China ' School of Health Services and Management, Hubei Preschool Teachers College, Wuhan, Hubei Province, China ' School of Health Services and Management, Hubei Preschool Teachers College, Wuhan, Hubei Province, China ' School of Health Services and Management, Hubei Preschool Teachers College, Wuhan, Hubei Province, China ' Hu Bei Rong Jun Hospital, Hongshan District Wuhan, Hubei, China

Abstract: The study aimed to assess a variety of Machine Learning tools for predicting treatment outcomes for patients based on the patient's electronic health record. The research framework is about supervised learning and ensemble methods, where algorithms classify the status of patient care learning from clinical data with their labels. Eight ML models were used for comparison, which include Decision Tree, Random Forest, Gradient Boosting, SVM, Gaussian Naive Bayes and Convolutional Neural Networks. Each model had been optimised by specific hyperparameter tuning, and their performance was assessed using metrics like mean absolute error, mean squared error, Matthew's correlation coefficient, F1-score and Area under the Curve of the Receiver Operating Characteristic. The evaluation metrics indicated that the best models are CNN, and AdaBoost. The comparative analysis serves as a reference for identifying apt ML methodologies in clinical decision support and opens a new path towards devising accurate and personalised treatment planning.

Keywords: predictions for patient treatment; machine learning algorithms; analysis of clinical data; comparative evaluations; performance assessments; clinical decision support; electronic health records.

DOI: 10.1504/IJIPT.2026.156311

International Journal of Internet Protocol Technology, 2026 Vol.19 No.3, pp.118 - 129

Received: 23 Apr 2025
Accepted: 02 Jun 2025
Published online: 10 Sep 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Comparative evaluation of machine learning-based algorithms for clinical treatment prediction using EHR data from China

Keep up-to-date