IDENTIFICATION OF DOMINANT FACTORS IN STUNTING CASE NARRATIVES USING PCA AND SVD APPROACHES

Debora Chrisinta (Universitas Timor, Indonesia)
Justin Eduardo Simarmata (Universitas Timor, Indonesia)
Milagros R. Baldemor (Don Mariano Marcos Memorial State University, Philippines)
Miko Purnomo (Universitas Timor, Indonesia)

Parameter: Jurnal Matematika, Statistika dan Terapannya
Vol 4 No 1 (2025): Parameter: Jurnal Matematika, Statistika dan Terapannya

Stunting remains a serious public health concern in Indonesia, exacerbated by the limited public understanding of its causes and prevention strategies. This study analyzes public perceptions of stunting based on reviews collected through web scraping from the 2023 Indonesian Health Survey (SKI). Text preprocessing techniques, Term Frequency-Inverse Document Frequency (TF-IDF) analysis, and dimensionality reduction using Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) were applied to a dataset comprising 21 reviews. The results indicate that PCA outperforms SVD in simplifying the relationships among key terms, as evidenced by a lower reconstruction error (0.003861 compared to 0.004232). The dominant factors influencing public perception include education, sanitation, and socio-economic conditions. These findings highlight the critical role of data-driven and visual-based educational strategies in enhancing public awareness and accelerating stunting prevention efforts.