Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Jun 16;13(12):2087.
doi: 10.3390/diagnostics13122087.

The Utilization of Machine Learning Algorithms for Assisting Physicians in the Diagnosis of Diabetes

Linh Phuong Nguyen  1   2 , Do Dinh Tung  2   3 , Duong Thanh Nguyen  4 , Hong Nhung Le  5 , Toan Quoc Tran  5 , Ta Van Binh  2 , Dung Thuy Nguyen Pham  6   7
Affiliations

The Utilization of Machine Learning Algorithms for Assisting Physicians in the Diagnosis of Diabetes

Linh Phuong Nguyen et al. Diagnostics (Basel). .

Abstract

This paper investigates the use of machine learning algorithms to aid medical professionals in the detection and risk assessment of diabetes. The research employed a dataset gathered from individuals with type 2 diabetes in Ninh Binh, Vietnam. A variety of classification algorithms, including Decision Tree Classifier, Logistic Regression, SVC, Ada Boost Classifier, Gradient Boosting Classifier, Random Forest Classifier, and K Neighbors Classifier, were utilized to identify the most suitable algorithm for the dataset. The results of the present study indicate that the Random Forest Classifier algorithm yielded the most promising results, exhibiting a cross-validation score of 0.998 and an accuracy rate of 100%. To further evaluate the effectiveness of the selected model, it was subjected to a testing phase involving a new dataset comprising 67 patients that had not been previously seen. The performance of the algorithm on this dataset resulted in an accuracy rate of 94%, especially the study's notable finding is the algorithm's accurate prediction of the probability of patients developing diabetes, as indicated by the class 1 (diabetes) probabilities. This innovative approach offers a meticulous and quantifiable method for diabetes detection and risk evaluation, showcasing the potential of machine learning algorithms in assisting clinicians with diagnosis and management. By communicating the diabetes score and probability estimates to patients, the comprehension of their disease status can be enhanced. This information empowers patients to make informed decisions and motivates them to adopt healthier lifestyle habits, ultimately playing a crucial role in impeding disease progression. The study underscores the significance of leveraging machine learning in healthcare to optimize patient care and improve long-term health outcomes.

Keywords: detection; diabetes; diabetes prediction; machine learning.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
The framework of machine learning.
Figure 2
Figure 2
Correlation heatmap.
Figure 3
Figure 3
Algorithms comparison for Train/Test splitting method.
Figure 4
Figure 4
Algorithms comparison using Train/Test splitting method.
Figure 5
Figure 5
Tuned Random Forest Classifier performance on predicting new data.
Figure 6
Figure 6
ROC curve of tuned Random Forest Classifier predicting new data.
See this image and copyright information in PMC

References

    1. International Diabetes Federation IDF Diabetes Atlas|Tenth Edition. [(accessed on 11 May 2023)]. Available online: https://diabetesatlas.org/
    1. World Health Organization The Growing Burden of Diabetes in Viet Nam. [(accessed on 11 May 2023)]. Available online: https://www.who.int/vietnam/news/feature-stories/detail/the-growing-burd....
    1. International Diabetes Federation Global Diabetes Data Report 2000–2045. [(accessed on 11 May 2023)]. Available online: https://diabetesatlas.org/data/
    1. International Diabetes Feferation Viet Nam Diabetes Report 2000–2045. [(accessed on 11 May 2023)]. Available online: https://diabetesatlas.org/data/en/country/217/vn.html.
    1. Russell S.J., Norvig P., Davis E. Artificial Intelligence: A Modern Approach. 3rd ed. Prentice Hall; Upper Saddle River, NJ, USA: 2010.