SEKARLANGIT, Sekarlangit and Widodo, Catur Edi and Tarno, Tarno (2024) MODEL HYBRID ARTIFICIAL NEURAL NETWORK DAN LONG SHORT-TERM MEMORY UNTUK PREDIKSI INDIKATOR PENCEMAR AIR SUNGAI. Masters thesis, UNIVERSITAS DIPONEGORO.

	Text COVER.pdf Download (933kB)
	Text BAB I.pdf Download (185kB)
	Text BAB II.pdf Download (621kB)
	Text BAB III.pdf Restricted to Repository staff only Download (452kB)
	Text BAB IV.pdf Restricted to Repository staff only Download (2MB)
	Text BAB V.pdf Download (145kB)
	Text DAFTAR PUSTAKA.pdf Download (200kB)
	Text LAMPIRAN.pdf Restricted to Repository staff only Download (520kB)

Abstract

Air sungai merupakan sumber daya yang sangat penting bagi kehidupan manusia dan ekosistem, namun sering kali terpapar pencemaran akibat aktivitas manusia seperti industri, limbah domestik, dan pertambangan. Pencemaran ini mengancam kulalitas air yang sangat dibutuhkan untuk keperluan manusia dan kehidupan organisme air. Identifikasi dini pencemaran air sungai menjadi penting, dan penggunaan teknologi informasi dengan pendekatan Artificial Intelligence muncul sebagai solusi potensial. Salah satu teknologi yang berkembang adalah peramalan atau prediksi indikator pencemar air sungai, yang memanfaatkan algoritma Artificial Intelligence berbasis time series yaitu, Artificial Neural Network (ANN) dan Long Short-Term Memory (LSTM). Model ANN dan LSTM menawarkan keunggulan dalam menggeneralisasi pola data serta kemampuan memodelkan dependensi temporal yang kompleks pada suatu data. Model prediksi Hybrid ANN-LSTM, yang dirancang dengan menggabungkan kedua pendekatan ini, digunakan untuk memperoleh prediksi yang lebih akurat terutama dalam kasus-kasus pencemaran air yang mempengaruhi indikator pencemar air sungai. Penelitian ini bertujuan untuk mengembangkan algoritma Hybrid ANN dan LSTM dalam memprediksi indikator pencemar air sungai, dengan harapan dapat meningkatkan efektivitas pemantauan pencemaran air sungai dan memberikan kontribusi dalam perlindungan lingkungan air.
Kata kunci: Artificial Neural Network (ANN), Model Hybrid, Long Short-Term Memory (LSTM), Prediksi, Indikator Pencemar Air Sungai

River water is a vital resource for ecosystems and human existence, but it is frequently contaminated by pollution from industries, mining, and household garbage. The quality of the water is at risk due to pollution, which is detrimental to both human health and the survival of aquatic life. Early detection of river water quality is crucial, and using Artificial Intelligence in information technology seems to be one way to address this issue. Forecasting or predicting the pollution indicator of river water is one of the emerging technologies. It makes use of time series-based Artificial Intelligence techniques, specifically Artificial Neural Network (ANN) and Long Short-Term Memory (LSTM). The capacity to represent intricate temporal correlations in data and the ability to generalize data patterns are two benefits of using ANN and LSTM models. The ANN-LSTM Hybrid prediction model, designed by combining these two approaches, is proposed to obtain more accurate predictions especially in cases of water pollution that affect river water pollution indicator. This research aims to develop a Hybrid ANN and LSTM algorithm in predicting river water pollution indicator, with the hope of increasing the effectiveness of monitoring river water pollution and contributing to the protection of the water environment.
Keywords: Artificial Neural Network (ANN), Hybrid Model, Long Short-Term Memory (LSTM), Prediction, River Water Pollution Indicator

Actions (login required)

View Item