TARIHORAN, DIKY ARIANTO (2025) ANALISIS KOMPARATIF ALGORITMA MACHINE LEARNING UNTUK KLASIFIKASI KELAYAKAN AIR TANAH BERDASARKAN PARAMETER FISIKA DI JAKARTA. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (499kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (52kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (352kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (128kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (682kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (30kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (112kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (1MB) |
Abstract
Groundwater quality is a critical factor in meeting clean water demands, particularly in areas such as Jakarta, which face increasing pressure from pollution caused by domestic waste, industrial activities, and septic tank seepage. This study aims to develop and evaluate groundwater feasibility classification models using nine machine learning algorithms based on physical water parameters. Data were collected in real-time from three administrative regions of Jakarta using Internet of Things (IoT) sensors that recorded pH, temperature, total dissolved solids (TDS), and turbidity. The final dataset consisted of 2,442 samples, which were split into 70% training and 30% testing sets to ensure model validity. Evaluation was conducted through hyperparameter tuning, cross-validation, feature importance analysis, model interpretation using LIME, and measurement of metrics including AUC, accuracy, precision, recall, and F1-score. Results showed that the CatBoost algorithm delivered the best overall performance (AUC 0.9448; accuracy 0.9318; F1-score 0.9209). LightGBM (AUC 0.9431; F1-score 0.9211) and XGBoost (AUC 0.9357; recall 0.9359) demonstrated competitive performance. Random Forest (precision 0.9094) and AdaBoost (recall 0.9327) yielded stable results, while Support Vector Machine (SVM) showed the lowest performance (AUC 0.8860; accuracy 0.8499). Based on the evaluation results, CatBoost is identified as the most ideal algorithm among those compared. Keywords: Groundwater Quality, Machine Learning, IoT, Classification Kualitas air tanah menjadi faktor krusial dalam pemenuhan kebutuhan air bersih, khususnya di wilayah seperti Jakarta yang menghadapi tekanan akibat pencemaran dari limbah domestik, aktivitas industri, dan rembesan septic tank. Penelitian ini bertujuan mengembangkan dan mengevaluasi model klasifikasi kelayakan air tanah menggunakan sembilan algoritma machine learning berbasis parameter fisika. Data dikumpulkan secara real-time dari tiga wilayah administratif Jakarta menggunakan sensor Internet of Things (IoT) yang merekam parameter pH, suhu, total dissolved solids (TDS), dan kekeruhan. Dataset akhir terdiri dari 2.442 sampel, yang kemudian dibagi menjadi 70% data pelatihan dan 30% data pengujian untuk memastikan validitas model. Evaluasi dilakukan melalui penyetelan hyperparameter, validasi silang, analisis feature importance, interpretasi menggunakan LIME, serta pengukuran metrik AUC, akurasi, presisi, recall, dan F1-score. Hasil menunjukkan bahwa algoritma CatBoost memberikan performa terbaik (AUC 0,9448; akurasi 0,9318; F1-score 0,9209). LightGBM (AUC 0,9431; F1-score 0,9211) dan XGBoost (AUC 0,9357; recall 0,9359) menunjukkan kinerja kompetitif. Random Forest (presisi 0,9094) dan AdaBoost (recall 0,9327) mencatat hasil yang stabil. Support Vector Machine (SVM) memiliki performa terendah (AUC 0,8860; akurasi 0,8499). Berdasarkan hasil perbandingan, algoritma CatBoost menjadi model yang paling ideal dibandingkan algoritma lainnya. Kata kunci: Kualitas Air Tanah, Machine Learning, IoT, Klasifikasi
Actions (login required)
 |
View Item |
