Klasifikasi Status Kesehatan dan Pola Morfologi Janin Menggunakan Metode Decision Tree dan Random Forest Berbasis Optimasi Fitur Genetic Algorithm

Afandi, Rizal (2026) Klasifikasi Status Kesehatan dan Pola Morfologi Janin Menggunakan Metode Decision Tree dan Random Forest Berbasis Optimasi Fitur Genetic Algorithm. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221116-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Kesehatan bayi selama dalam kandungan merupakan hal penting yang harus dipantau untuk mencegah bahaya gawat janin. Salah satu alat yang sering digunakan oleh dokter adalah Cardiotocography (CTG) untuk merekam detak jantung bayi dan kontraksi rahim ibu. Namun, data dari alat ini cukup rumit untuk dibaca secara manual karena banyaknya informasi yang terekam. Penelitian ini bertujuan untuk membuat sistem otomatis yang bisa mengelompokkan kondisi kesehatan janin (3 kelas) dan bentuk pola morfologinya (10 kelas) menggunakan metode Decision Tree dan Random Forest. Mengingat adanya beberapa permasalahan pada dataset, Algoritma Genetika (GA) diimplementasikan untuk menyeleksi fitur secara evolusioner guna membuang fitur noise, yang kemudian diikuti dengan penerapan metode Synthetic Minority Over-sampling Technique for Nominal and Continuous (SMOTE-NC) pada data latih untuk menangani masalah ketimpangan data dan menciptakan keseimbangan distribusi antar kelas. Hasil penelitian menunjukkan bahwa arsitektur Random Forest dengan integrasi Algoritma Genetika dan SMOTE-NC (RF-GA-SMOTE-NC) merupakan model terbaik dengan perolehan nilai AUC Score, Accuracy, dan Macro F1-Score tertinggi dibandingkan 5 model lainnya, yaitu DT, RF, DT-GA, RF-GA dan DT-GA-SMOTE-NC. Pada klasifikasi 3 kelas, model ini memperoleh tingkat pengujian dengan AUC Score sebesar 98,69%, Accuracy 95,04%, Macro Precision 92,17%, Macro Recall 91,25%, dan Macro F1-Score 91,50%. Sementara pada klasifikasi 10 kelas, model mencapai AUC Score 98,84%, Accuracy 85,11%, Macro Precision 78,95%, Macro Recall 82,09%, dan Macro F1-Score 80,24%. Hasil dari penelitian ini diharapkan dapat membantu tenaga medis dalam mendeteksi masalah kesehatan janin secara lebih cepat, tepat, dan objektif.
=======================================================================================================================================
Fetal health during pregnancy is an important factor that must be monitored to prevent fetal distress. One of the tools often used by doctors is Cardiotocography (CTG) to record the baby's heart rate and the mother's uterine contractions. However, the data from this tool is quite complicated to read manually due to the large amount of information recorded. This study aims to create an automated system that can classify fetal health conditions (3 classes) and morphological patterns (10 classes) using the Decision Tree and Random Forest methods. Given several issues present in the dataset, a Genetic Algorithm (GA) is implemented for evolutionary feature selection to remove noise attributes, followed by the application of the Synthetic Minority Over-sampling Technique for Nominal and Continuous (SMOTE-NC) on the training data to address data imbalance and achieve a balanced class distribution. The results indicate that the Random Forest architecture with the integration of GA and SMOTE-NC (RF-GA-SMOTE-NC) is the best model, achieving the highest AUC Score, Accuracy, and Macro F1-Score compared to 5 other models, namely DT, RF, DT-GA, RF-GA, and DT-GA-SMOTE-NC. In the 3-class classification, this model obtained a testing AUC Score of 98.69%, Accuracy of 95,04%, Macro Precision of 92,17%, Macro Recall of 91,25%, and Macro F1-Score of 91,50%. Meanwhile, in the 10-class classification, the model achieved an AUC Score of 98.84%, Accuracy of 85.11%, Macro Precision of 78.95%, Macro Recall of 82.09%, and Macro F1-Score of 80.24%. The results of this study are expected to help medical personnel detect fetal health problems more quickly, accurately, and objectively.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Algoritma Genetika, Kesehatan Janin, Random Forest, SMOTE-NC, Decision Tree, Fetal Health, Genetic Algorithm
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > Q Science (General) > Q370 Entropy (Information theory) Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Rizal Afandi
Date Deposited:	30 Jul 2026 01:31
Last Modified:	30 Jul 2026 15:09
URI:	http://repository.its.ac.id/id/eprint/139202

Actions (login required)

View Item