Data_Stunting

Citation Author(s):
dicky
faisal
allya
paramita
Submitted by:
dicky zein
Last updated:
Tue, 08/08/2023 - 01:50
DOI:
10.21227/7wx2-3c69
Data Format:
License:
58 Views
Categories:
Keywords:
0
0 ratings - Please login to submit your rating.

Abstract 

ommon approaches to stunting prediction, including statistical analysis and machine learning, have poor performance due to shifts in the factors influencing stunting. Causes data cannot be integrated directly when using statistical analysis. At the same time, machine learning causes a decrease in predictive performance down over time. This study proposes a new approach to stunting prediction in infants and toddlers aged 0-5 years using continuous learning methods. The method used uses the PackNet algorithm. Each sample in the India Demographic and Health Survey 2019-2021 (IDHS) data is grouped into potentially normal and potentially stunting. Data with a significant variable on stunting is selected according to the literature and logistic regression test with a significance value of 5% using SPSS as data input. Based on experiments conducted using data from IDHS models trained by the PackNet algorithm showed better performance than those trained with Artificial Neural Networks.The proposed method performs better when learning from IDHS 2019-2021 data and can adapt well to new tasks.

Instructions: 

The data in the row shows the sample and in the column shows the variables that affect stunting.

Comments

.

Submitted by jean-pierre morard on Tue, 08/08/2023 - 08:26