Machine Learning untuk Deteksi Cemaran Mikroplastik pada Air Berdasarkan Data Electronic Tongue

Putri, Rizqy Ahsana (2025) Machine Learning untuk Deteksi Cemaran Mikroplastik pada Air Berdasarkan Data Electronic Tongue. Masters thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 6025231069-Master_Thesis.pdf] Text
6025231069-Master_Thesis.pdf
Restricted to Repository staff only until 1 April 2027.

Download (5MB) | Request a copy

Abstract

Pencemaran mikroplastik terutama di lingkungan air telah menjadi masalah global yang mempengaruhi kelestarian lingkungan dan kesehatan manusia. Untuk mengatasi masalah ini, dibutuhkan metode deteksi yang efisien dan akurat untuk mengidentifikasi keberadaan cemaran mikroplastik dalam air. Metode yang selama ini digunakan untuk deteksi mikroplastik, seperti Infrared Imaging Microscope, memiliki beberapa keterbatasan, antara lain biaya yang tinggi, waktu analisis yang lama, serta kebutuhan akan keahlian khusus dalam pengoperasiannya. Alternatif yang lebih efisien adalah penggunaan voltammetri yang dikombinasikan dengan machine learning, karena metode ini lebih cepat, murah, dan mudah dioperasikan dibandingkan pendekatan konvensional. Oleh karena itu, penelitian ini bertujuan untuk mengusulkan pendekatan baru menggunakan machine learning berdasarkan data voltammetri untuk mengestimasi kandungan cemaran mikroplastik dalam air. Metodologi penelitian dimulai dengan tahap ekstraksi fitur. Selanjutnya, tiap dataset hasil ekstraksi akan diproses menggunakan metode regresi. Terdapat lima metode yang diterapkan, yaitu: (1) DNN, (2) GRU, (3) Transformer Encoder, (4) SVR, dan (5) Regresi Linier. Kinerja model klasifikasi akan dievaluasi menggunakan metrik R-squared (R²), Root Mean Squared Error (RMSE), dan Mean Squared Error (MSE). Hasil penelitian menunjukkan bahwa DNN unggul di model konsentrasi sedang (100 ppm – 500 ppm) maupun di model konsentrasi tinggi (600 ppm – 1000 ppm). Evaluasi DNN pada model sedang mencapai R² sebesar 0.84 dan RMSE 55.55, sedangkan model regresi untuk konsentrasi tinggi mencapai R² sebesar 0.97 dengan RMSE 22.47.
=====================================================================================================================================
Microplastic pollution, especially in the aquatic environment, has become a global problem that affects environmental sustainability and human health. To overcome this problem, an efficient and accurate detection method is needed to identify the presence of microplastic contamination in water. Methods that have been used for microplastic detection, such as Infrared Imaging Microscope, have several limitations, including high cost, long analysis time, and the need for special expertise in its operation. A more efficient alternative is the use of voltammetry combined with machine learning, because this method is faster, cheaper, and easier to operate than conventional approaches. Therefore, this study aims to propose a new approach using machine learning based on voltammetry data to estimate the content of microplastic contamination in water. The research methodology starts with the feature extraction stage. Next, each extracted dataset will be processed using regression methods. There are five methods applied, namely: (1) DNN, (2) GRU, (3) Transformer Encoder, (4) SVR, and (5) Regresi Linier. The performance of the classification model will be evaluated using the R-squared (R²), Root Mean Squared Error (RMSE), and Mean Squared Error (MSE) metrics. The results showed that DNN excelled in the medium concentration model (100 ppm - 500 ppm) as well as in the high concentration model (600 ppm - 1000 ppm). The DNN evaluation in the medium model achieved an R² of 0.84 and an RMSE of 55.55, while the regression model for medium concentration achieved an R² of 0.97 with an RMSE of 22.47.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Electronic Tongue, Machine learning, Microplastic, Mikroplastik, Regresi, Regression Analysis, Voltammetri, Voltammetry.
Subjects: T Technology > T Technology (General) > T57.5 Data Processing
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User: Rizqy Ahsana Putri
Date Deposited: 06 Feb 2025 06:27
Last Modified: 06 Feb 2025 06:27
URI: http://repository.its.ac.id/id/eprint/118475

Actions (login required)

View Item View Item