Claim Missing Document
Check
Articles

Found 1 Documents
Search

LSTM Parameter Optimization with Genetic Algorithm for Stunting Prediction Rifqi Muhammad Fikri; Ayi Purbasari; Arief Zulianto; Fedri Ruluwedrata Rinawan; Ari Indra Susanti
PIKSEL : Penelitian Ilmu Komputer Sistem Embedded and Logic Vol. 13 No. 1 (2025): Maret 2025
Publisher : LPPM Universitas Islam 45 Bekasi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33558/piksel.v13i1.10761

Abstract

Stunting is caused by a lack of nutrients or sickness, and stunted children may have impaired immune systems, increased mortality rates, and are more prone to endure long-term developmental abnormalities. Stunting prevalence in Indonesia remains concerningly high by the end of 2024. Through the use of integrated health posts, or pos pelayanan terpandu (Posyandu), and the technology-based website iPosyandu, attempts are being made to lower the prevalence of stunting. Using toddler data from iPosyandu, this study proposes a Long-Short Term Memory (LSTM) model for predicting stunting based on WHO standards, categorizing children as tall, normal, stunted, or severly stunted. By using a genetic algorithm (GA) for learning rate hyperparameter tuning, the LSTM model is significantly improved. Five generations, each with five populations, were used for the GA-based optimization, which explored learning rates ranging from 5.23E-04 to 8.83E-03. The results show that 7.82E-03 was the optimal learning rate, producing the greatest accuracy of 91.10%, indicating that this range improves the performance of LSTM models.