Early diabetes risk prediction using Ant Colony Optimization algorithm / Nur Aisyatul Husna Ahmad Yusri and Rizauddin Saian

Ahmad Yusri, Nur Aisyatul Husna and Saian, Rizauddin (2023) Early diabetes risk prediction using Ant Colony Optimization algorithm / Nur Aisyatul Husna Ahmad Yusri and Rizauddin Saian. In: Research Exhibition in Mathematics and Computer Sciences (REMACS 5.0). College of Computing, Informatics and Media, UiTM Perlis, pp. 159-160. ISBN 978-629-97934-0-3
Abstract

Diabetes is a deadly disease that causes serious health complications to its sufferers. It costs the sufferers' health as well as their money. It is crucial to detect diabetes risk early to prevent the disease from worsening and becoming hard to treat. Therefore, this study has developed a classification model for predicting early diabetes risk using an Ant Colony Optimization (ACO) algorithm. The ACO-based classification algorithm, Ant-Miner is used to train the diabetes dataset of 520 new diabetes or potential diabetes patients from Sylhet Diabetes Hospital in Sylhet, Bangladesh. The average predictive accuracy from Ant-Miner is compared to the average predictive accuracy from J48. It is found that the average predictive accuracy of the model produced by Ant-Miner is at par with J48. The average predictive accuracy of the model produced by Ant-Miner is 95.51%, while J48 is 95.38%.

Item Details
Edit Item
Edit Item
Downloads & Files
[thumbnail of 100257.pdf]
Text
100257.pdf
Download (1MB)
Indexing & Metrics