Perancangan Dan Implementasi Model Fuzzy-Naive Bayes untuk Klasifikasi Kualitas Air Bersih

Rahmi, Luthfiah Khafifah (2026) Perancangan Dan Implementasi Model Fuzzy-Naive Bayes untuk Klasifikasi Kualitas Air Bersih. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221077-Undergraduate_Thesis1.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Penelitian ini mengusulkan perancangan dan implementasi arsitektur hibrida Fuzzy-Naïve Bayes untuk meningkatkan stabilitas dan akurasi klasifikasi mutu air secara real-time. Sistem akuisisi data dibangun menggunakan mikrokontroler ESP32 yang terintegrasi dengan sensor pH (pH-4502C), Total Dissolved Solids (TDS DFRobot SEN0244), suhu (DS18B20), serta modul potensiostat elektrokimia untuk deteksi ion logam Besi (Fe) dan Tembaga (Cu). Data sensor mentah diproses terlebih dahulu melalui Fuzzy Inference System (FIS) metode Mamdani dengan fungsi keanggotaan trapesium dan segitiga, lalu didefuzzifikasi menggunakan metode Centroid untuk menghasilkan Fuzzy Quality Index (FQI). Skor FQI tersebut kemudian diumpankan sebagai fitur tunggal ke dalam algoritma Gaussian Naïve Bayes yang dieksekusi pada server backend FastAPI dan di-host pada platform Hugging Face Spaces. Hasil pengujian menunjukkan bahwa klasifikasi Naïve Bayes menggunakan data mentah tanpa ekstraksi fitur hanya mencapai akurasi sebesar 62,77% (5-fold CV 59,50%). Sebaliknya, penerapan arsitektur hibrida FQI + Naïve Bayes berhasil meningkatkan akurasi data uji secara signifikan hingga 81,33% (5-fold CV 80,90%) pada dataset sekunder, serta mencapai akurasi 95,33% pada pengujian dataset primer sensor lapangan. Pengujian real-time pada sampel air keran, larutan garam, dan air sungai Mulyosari membuktikan bahwa sistem hibrida ini memiliki ketahanan tinggi terhadap noise sensorik dengan tingkat keyakinan prediksi hingga 96,34%.
=====================================================================================================================================
This study proposes the design and implementation of a hybrid Fuzzy-Naïve Bayes architecture to enhance the stability and accuracy of real-time water quality classification. The data acquisition hardware utilizes an ESP32 microcontroller integrated with a pH sensor (pH-4502C), Total Dissolved Solids sensor (TDS DFRobot SEN0244), temperature sensor (DS18B20), and an electrochemical potentiostat module for detecting Iron (Fe) and Copper (Cu) metal ions. Raw sensor data are pre-processed via a Mamdani Fuzzy Inference System (FIS) employing trapezoidal and triangular membership functions, followed by Centroid defuzzification to generate a Fuzzy Quality Index (FQI). This FQI score is subsequently fed as a single feature input into a Gaussian Naïve Bayes classifier deployed on a FastAPI backend hosted on Hugging Face Spaces. Experimental results indicate that direct Naïve Bayes classification on raw multi-sensor data yields a low-test accuracy of 62.77% (5-fold CV 59.50%). In contrast, the proposed hybrid FQI + Naïve Bayes architecture significantly improves test accuracy to 81.33% (5-fold CV 80.90%) on secondary datasets and achieves up to 95.33% accuracy on primary field sensor datasets. Real-time implementation on tap water, saline solutions, and river water samples demonstrates high robustness against sensory noise, obtaining prediction confidence scores up to 96.34%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Water Quality Monitoring, IoT, ESP32, Fuzzy Logic, Naïve Bayes Classifier, Water Quality Monitoring, IoT, ESP32, Fuzzy Logic, Naïve Bayes, Centroid, Classification.
Subjects:	Q Science > QA Mathematics > QA9.64 Fuzzy logic T Technology > TD Environmental technology. Sanitary engineering > TD259.2 Drinking water. Water quality T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK351 Electric measurements. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Luthfiah Khafifah Rahmi
Date Deposited:	20 Aug 2026 01:36
Last Modified:	20 Aug 2026 01:36
URI:	http://repository.its.ac.id/id/eprint/144344

Actions (login required)

View Item