Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2024
- Goals
- Record
- c3fc7bd5-bef0-4b60-bcf6-df0b0271a70b
- Published
- 2026-10-30
UDINUS conducted a research activity on early diabetes prediction by applying a Decision Tree Learning Algorithm and integrating the Adaboost Method. The activity participants were Alam, Laskar; Fanani, Ahmad Zainul; and Affandy. The activity addressed the importance of early diabetes prediction and efforts to increase prediction accuracy as part of the stated goal concerning diabetes prediction and prevention of death.
The activity used a dataset from Kaggle containing 520 records, 16 attributes, and one positive or negative diabetes class. The Decision Tree Learning Algorithm was applied as the predictive approach, and the Adaboost Method was integrated to increase the accuracy of the prediction. Evaluation was conducted using a Confusion Matrix. The Decision Tree algorithm achieved an accuracy of 94.23%. Following integration of the Adaboost Method, the reported accuracy increased to 97.31%.
The findings emphasize the importance of predictive approaches in early disease detection. They also highlight the potential of the Adaboost Method to improve the accuracy of diabetes prediction. In this activity, the comparison between the Decision Tree algorithm and the integrated Adaboost approach provides the stated basis for describing the improvement in prediction accuracy. The activity therefore presents a health-related research contribution focused on an early diabetes prediction method, its evaluation, and the reported implications of the findings. This evidence draft records UDINUS's activity and the participation of the named researchers within the scope described in the source.