Machine Learning-Based Cancer Risk Prediction Using Clinical And Lifestyle Factors
IOSR Journal of Computer Engineering · 2026-02 · 28 (2) · 29-35
DOI: 10.9790/0661-2802012935 ↗Kurz gefasst
Fachartikel in der Zeitschrift IOSR Journal of Computer Engineering, 2026-02. Verfasst von Thi Hong Loan Nguyen.
Abstract (Originalsprache)
Background
Cancer has become one of the leading causes of mortality worldwide, with increasing incidence and a noticeable shift toward younger age groups in many countries. Early identification of cancer risk is therefore essential for improving prevention strategies and supporting timely medical intervention. Recent advances in machine learning have enabled the analysis of clinical and lifestyle data to identify potential disease risks and support decision-making in healthcare. However, selecting an appropriate predictive model remains an important challenge in developing reliable prediction systems. Therefore, this study aims to develop a machine learning–based approach for cancer risk prediction using clinical and lifestyle factors and to determine the most effective predictive model through comparative analysis. Materials and Methods: In this study, a dataset containing several clinical and lifestyle risk indicators was used, including age, gender, body mass index (BMI), smoking status, genetic risk, physical activity level, alcohol consumption, and cancer history. The dataset was divided into training and testing sets using stratified sampling to preserve class distribution. Several machine learning algorithms were implemented and compared, including Logistic Regression, Decision Tree, Random Forest, Support Vector Machine (SVM), k-Nearest Neighbors (KNN), and AdaBoost. Hyperparameter optimization was performed using GridSearch combined with stratified cross-validation to improve model performance and ensure reliable evaluation.
Results
The experimental results show that the evaluated machine learning models are capable of predicting cancer risk using clinical and lifestyle features. Among the tested algorithms, the AdaBoost model achieved the best predictive performance with an accuracy of 0.96 and a ROC-AUC of approximately 0.97. Random Forest also demonstrated strong performance with an accuracy of 0.94 and a ROC-AUC of about 0.96. Other models such as KNN, SVM, Decision Tree, and Logistic Regression showed comparatively lower predictive performance.
Conclusion
The results indicate that the AdaBoost-based prediction approach can effectively identify potential cancer risk using clinical and lifestyle data. This method has potential application in supporting early cancer risk assessment and assisting the development of data-driven decision support systems in healthcare.
Zitieren
APA
Nguyen, T. H. L. (2026). Machine Learning-Based Cancer Risk Prediction Using Clinical And Lifestyle Factors. IOSR Journal of Computer Engineering, 28(2), 29-35. https://doi.org/10.9790/0661-2802012935
Vancouver
Nguyen THL. Machine Learning-Based Cancer Risk Prediction Using Clinical And Lifestyle Factors. IOSR Journal of Computer Engineering. 2026;28(2):29-35. doi:10.9790/0661-2802012935
BibTeX
@article{Nguyen2026,
author = {Nguyen, Thi Hong Loan},
title = {Machine Learning-Based Cancer Risk Prediction Using Clinical And Lifestyle Factors},
journal = {IOSR Journal of Computer Engineering},
year = {2026},
volume = {28},
number = {2},
pages = {29--35},
doi = {10.9790/0661-2802012935}
}
EUVIMED