PRATAMA, IRGY AHMAD RIZKI (2025) ANALISIS PREDIKSI CUACA BERBASIS MACHINE LEARNING DENGAN LSTM (Studi Kasus: Kota Tangerang). S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (527kB) | 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 (263kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (364kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (368kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (1MB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (253kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Repository staff only Download (216kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (594kB) |
Abstract
Increasingly unpredictable weather changes require an accurate and responsive weather prediction system. Weather prediction plays an important role in decision making in various sectors such as agriculture, transportation, construction, and disaster mitigation. However, conventional methods such as linear regression and autoregressive statistical models are often unable to capture the complexity and temporal dynamics of weather data. Therefore, this study applies machine learning, namely Long Short-Term Memory (LSTM), which has proven to be superior in modeling time series data that has long-term dependencies. This study focuses on daily weather predictions for the next seven days in Tangerang City. The data used comes from the OpenWeatherMap API, including six main parameters: temperature, humidity, air pressure, wind speed, rainfall, and solar radiation. The preprocessing process involves data normalization and conversion to time series format. The LSTM model is built with two hidden layers and the EarlyStopping technique to avoid overfitting, then trained using a historical dataset that is divided proportionally (80:20). The model prediction results are evaluated using three main metrics, namely Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). The experimental results show that the LSTM model is able to predict three main weather parameters (temperature, humidity, and air pressure) with a relatively low and stable error rate. This study shows that LSTM is an effective and adaptive solution in modeling short-term weather patterns based on real-time data. Keywords: artificial intelligence, LSTM, weather prediction, real-time data, Tangerang City Perubahan cuaca yang semakin tidak menentu menuntut adanya sistem prediksi cuaca yang akurat dan responsif. Prediksi cuaca memainkan peran penting dalam pengambilan keputusan di berbagai sektor seperti pertanian, transportasi, konstruksi, dan mitigasi bencana. Namun, metode konvensional seperti regresi linier dan model statistik autoregressive seringkali tidak mampu menangkap kompleksitas dan dinamika temporal data cuaca. Oleh karena itu, penelitian ini menerapkan machine learning, yaitu Long Short-Term Memory (LSTM), yang terbukti unggul dalam pemodelan data deret waktu (time series) yang memiliki dependensi jangka panjang. Studi ini fokus pada prediksi cuaca harian selama tujuh hari ke depan di Kota Tangerang. Data yang digunakan bersumber dari API OpenWeatherMap, meliputi enam parameter utama: temperatur, kelembapan, tekanan udara, kecepatan angin, curah hujan, dan radiasi matahari. Proses preprocessing melibatkan normalisasi data serta konversi ke format time series. Model LSTM dibangun dengan dua lapisan tersembunyi dan teknik EarlyStopping untuk menghindari overfitting, kemudian dilatih menggunakan dataset historis yang dibagi secara proporsional (80:20). Hasil prediksi model dievaluasi menggunakan tiga metrik utama, yaitu Mean Absolute Error (MAE), Root Mean Square Error (RMSE), dan Mean Absolute Percentage Error (MAPE). Hasil eksperimen menunjukkan bahwa model LSTM mampu memprediksi tiga parameter utama cuaca (temperatur, kelembapan, dan tekanan udara) dengan tingkat kesalahan yang masih tergolong rendah dan stabil. Penelitian ini menunjukkan bahwa LSTM merupakan solusi yang efektif dan adaptif dalam memodelkan pola cuaca jangka pendek berbasis data real-time. Kata kunci: Machine Learning, LSTM, Prediksi cuaca, Data Real-time, Kota Tangerang.
Actions (login required)
 |
View Item |
