Skripsi

PENERAPAN METODE LONG SHORT-TERM MEMORY (LSTM) UNTUK PREDIKSI PRODUKSI MINYAK BUMI BERDASARKAN DATA TIME SERIES HARIAN

---

Oil production prediction is a crucial component of operational planning and strategic decision-making in the upstream oil and gas industry. This study applies the Long Short-Term Memory (LSTM) method to model and predict oil production using daily historical data from PT Pertamina Hulu Rokan Regional 1 Zona 4 Limau Field for the period January 1, 2022, to July 31, 2025. Data processing steps include data cleaning, normalization using Min–Max Scaling, data partitioning, method evaluation, and sequence generation based on a 30-day window size to capture long-term temporal dependencies. The LSTM method was built using a single-layer LSTM architecture containing 64 neuron units, a dropout rate of 0.2 to mitigate overfitting, and a dense output layer with one neuron. The training process was performed using the Adam optimization algorithm and the Mean Squared Error (MSE) loss function. Performance evaluation yielded a Root Mean Square Error (RMSE) of 100,49 and a Mean Absolute Percentage Error (MAPE) of 1.72%. This indicates a low prediction error rate and the method's ability to accurately represent temporal patterns. The trained method was then used to predict production for the next three years (2025–2028). The prediction results showed a stable pattern consistent with historical trends, confirming the effectiveness of the LSTM approach in modeling time series oil production data.

---

Availability

Detail Information

Series Title

-

Call Number

T1895542025

Publisher

Indralaya : Prodi Sistem Informasi, Fakultas Ilmu Komputer Universitas Sriwijaya., 2025

Collation

xv, 76 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

519.550 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Sistem Informasi
Analisis Data Time Series

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment