• DOI:10.4108/EAI.19-10-2018.2281287
• Corpus ID: 140121665

Decision Tree Algorithm Using Particle Swarm Optimization To Improve The Accuracy Of Detection Malnutrition In Toddler

@article{Agustina2019DecisionTA,
  title={Decision Tree Algorithm Using Particle Swarm Optimization To Improve The Accuracy Of Detection Malnutrition In Toddler},
  author={Candra Agustina and Nani Purwati and Gunawan Budi Sulistyo and Noor Hasan and Paulus Rapiyanta},
  journal={Proceedings of the International Conference of Science and Technology for the Internet of Things},
  year={2019}
}

• C. Agustina, Nani Purwati, +2 authors Paulus Rapiyanta
• Published 14 April 2019
• Computer Science
• Proceedings of the International Conference of Science and Technology for the Internet of Things

Malnutrition in Indonesia is still relatively high, it is recorded that there are 19.6% of children under five years old who suffer from malnutrition throughout Indonesia. Malnutrition will give impacts tochildren’s health in the future. Therefore, the action to detect the malnutrition occurrence should be conducted as early as possible; thus, the patient will immediately getthe right health care. Many methods have already implemented to determine whether a toddler suffers from malnutrition or not. One of them is by using data mining techniques to create a grouping. Toddlers will be categorized into 4 groups namely Good Nutrition, Lack of Nutrition, Over Nutrition and Malnutrition. The data used are Toddler data, which is consisted of 4 predictor attributes and 1 result attribute. In the previous research the algorithm used was C 4.5 that was compared toBack-propagation. The result of the data processing by using C 4.5 algorithm is 88.24% and Kappa with the amount of 0.725. In order to improve the accuracy of the C 4.5 algorithm, the algorithm of Particle Swarm Optimization is implementedfor the optimization. Having implemented Particle Swarm Optimization, the accuracy is obtained in the amount of 98.04% and Kappa 0.954. Accordingly, the Particle Swarm Optimization increases the accuracy of C 4.5 by 9.80%. The feature selection, which is conducted, indicates that the attribute of family status must be omitted to obtain higher amount of accuracy.