Data Mining Performance of Toddler Nutrition Classification Based on Family Nutrition Awareness and Human Development Index
Darmansyah1, Gede Putra Kusuma2
1Darmansyah, Computer Science Department, BINUS Graduate Program – Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia.
2Gede Putra Kusuma*, Computer Science Department, BINUS Graduate Program – Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia.
Manuscript received on January 02, 2020. | Revised Manuscript received on January 15, 2020. | Manuscript published on January 30, 2020. | PP: 1591-1596 | Volume-8 Issue-5, January 2020. | Retrieval Number: E4573018520/2020©BEIESP | DOI: 10.35940/ijrte.E4573.018520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Nutrition problems that occurred in districts/cities of Central Java province from 2015-2017 were only 1 district city that did not have nutritional problems (good category) in 2015.The rest had acute, chronic or acute chronic nutrition problems. The search for the most influential attributes in toddler nutrition problems using data mining is expected to help health workers to focus more on solving problems based on classification in the area.Therefore, improving the nutritional status of the community can be accelerated. The best parameter search from the selection of features and data mining algorithm using the Optimize Parameters (Grid) operator found in Rapidminer.The feature selection models used are Backward Elimination, Forward Selection, and Optimize Selection. The datamining algorithm used is Naive Bayes, Decision Tree, k-NN, and Neural Network.The merging of the feature selection model and the datamining algorithm resulted in 12 algorithm models used in this study.The best model that was processed using test data with the highest accuracy of 74.19% was obtained from backward-neural network elimination. The attribute that is not very influential based on the model obtained is the condition of the mother who died.
Keywords: Feature Selection, Data Mining, Parameter Tuning.
Scope of the Article: Data Mining.