Maharani, Syarifah Kusuma (2024) Perbandingan Metode Machine Learning Dalam Klasifikasi Potensi Keluarga Beresiko Stunting Pada Kecamatan Merbau. Other thesis, Universitas Islam Riau.

Text 193510512.pdf - Submitted Version Download (5MB)

Abstract

Stunting is a condition where chronic malnutrition is caused by insufficient nutritional intake over a long period of time due to administration eating food that doesn't meet your needs. In Indonesia, according to the results of the Infant Nutrition Survey Indonesia (SSGBI) stunting prevalence will reach 21.6% in 2022. In this case, The family also participates in monitoring, checking the risk status of the family The risk of stunting still takes quite a long time because it is done routinely Manuals are also prone to inaccuracies. Therefore, to classify families with the potential for stunting can be done using a machine algorithm learning, namely K-Nearest Neighbor, Decision Tree, Naïve Bayes, and Random Forest. The dataset used was taken from 3 types of data, namely PK, KB and KPD data. Testing was carried out by applying the K-fold cross validation method for eliminate bias in the data. The highest accuracy results are in K-fold cross validation namely the Decision Tree algorithm gets a score of 99.76% with a precision of 99.73%, recall 99.58%, and f-1score 99.65%. In the Random Forest algorithm we get accuracy value 99.25%. Then the K-NN algorithm gets 92.42% and The lowest accuracy is the Naïve Bayes algorithm with a value of 41.69%. In research This algorithm that has the best performance is the Deep Decision Tree algorithm classifying families at risk of stunting in Merbau District

Actions (login required)

View Item