Pembuatan Model Machine Learning Untuk Prediksi Nilai Total Organic Carbon (Toc) Pada Sistem Pengolahan Air: Studi Kasus Pt. Finusolprima Farma Internasional

Mustapa, Dieki Rian (2024) Pembuatan Model Machine Learning Untuk Prediksi Nilai Total Organic Carbon (Toc) Pada Sistem Pengolahan Air: Studi Kasus Pt. Finusolprima Farma Internasional. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6047221036-Master_Tesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (3MB) | Request a copy

Abstract

Kualitas air sangat penting dalam produksi farmasi, di mana air berfungsi sebagai pelarut dan bahan baku. Kontaminasi dengan senyawa organik menimbulkan risiko terhadap integritas dan keamanan produk. TOC berfungsi sebagai indikator utama untuk menilai tingkat polusi organik dalam air. Penelitian ini bertujuan untuk membandingkan tiga algoritma machine learning yang berbeda - Regresi Linear (RL), Random Forest (RF), dan multilayer perceptron (MLP) - untuk memprediksi Total Organic Carbon (TOC) dalam sistem pengolahan air farmasi dan penerapannya dalam industri farmasi. Dengan memanfaatkan data train yang mencakup berbagai kondisi operasional sistem pengolahan air farmasi, penelitian ini melakukan analisis yang komprehensif. Setiap algoritma menjalani evaluasi menggunakan metrik kinerja seperti mean squared error (MSE), koefisien determinasi (R-squared), dan akurasi prediksi untuk menilai efektivitas mereka dalam memprediksi konsentrasi TOC. Studi ini menemukan bahwa hasil koefisien korelasi secara berurutan dari tertinggi ke terendah adalah model RF, MLP, dan RL dengan nilai masing-masing sebesar 95,04%, 93,11%, dan 80,27%. Demikian pula, parameter evaluasi metrik lainnya seperti Root mean squared error (RMSE), Mean absolute error (MAE), Root relative squared error (RRSE), dan Relative absolute error (RAE), memiliki nilai terendah hingga tertinggi pada model RF, MLP, dan RL. Penelitian ini menawarkan wawasan berharga dalam memanfaatkan algoritma machine learning untuk prediksi TOC dalam pengolahan air farmasi. Memahami kelebihan dan keterbatasan masing-masing algoritma memberdayakan industri farmasi untuk mengambil keputusan yang berbasis informasi, dan penerapan machine learning dalam industri farmasi.
========================================================================================================================================
Water quality is critical in pharmaceutical production, where water serves as a solvent and raw material. Contamination with organic compounds poses a risk to product integrity and safety. TOC serves as a key indicator to assess the level of organic pollution in water. This study aims to compare three different machine learning algorithms - Linear Regression (RL), Random Forest (RF), and multilayer perceptron (MLP) - for predicting Total Organic Carbon (TOC) in pharmaceutical water treatment systems and their applicability in the pharmaceutical industry. Utilizing a data train covering a wide range of operational conditions of pharmaceutical water treatment systems, this study conducted a comprehensive analysis. Each algorithm underwent evaluation using performance metrics such as mean squared error (MSE), coefficient of determination (R-squared), and prediction accuracy to assess their effectiveness in predicting TOC concentration. The study found that the correlation coefficient results in order from highest to lowest were RF, MLP, and RL models with values of 95.04%, 93.11%, and 80.27%, respectively. Similarly, other metric evaluation parameters such as Root mean squared error (RMSE), Mean absolute error (MAE), Root relative squared error (RRSE), and Relative absolute error (RAE), have the lowest to highest values in RF, MLP, and RL models. This research offers valuable insights in utilizing machine learning algorithms for TOC prediction in pharmaceutical water treatment. Understanding the advantages and limitations of each algorithm empowers the pharmaceutical industry to make informed decisions, and the application of machine learning in the pharmaceutical industry.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Machine learning;Data acquisition; Pengolahan air; Total Organic Carbon (TOC); Supervised Learning. ======================================================================================================================== Machine learning; Data acquisition; Water treatment; Total Organic Carbon (TOC); Supervised Learning.
Subjects:	T Technology > T Technology (General) > T174 Technological forecasting
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 78201-System And Technology Innovation
Depositing User:	Dieki Rian Mustapa
Date Deposited:	29 Jul 2024 13:41
Last Modified:	29 Jul 2024 13:41
URI:	http://repository.its.ac.id/id/eprint/109379

Actions (login required)

View Item