CD Skripsi

Implementasi Algoritma Cnn-Bilstm Untuk Memprediksi Curah Hujan Sebagai Pertimbangan Masa Penanaman Cabai

---

ABSTRACT
The fluctuating rainfall pattern in Pekanbaru poses a challenge in determining the optimal planting time for red chili cultivation. Reliance on traditional planting calendars often increases the risk of crop failure since it does not account for climate variability. This study develops a daily rainfall prediction model using the CNN–BiLSTM architecture with BMKG weather data from January 2014 to June 2025. The dataset was processed through interpolation, time-series transformation, moving averages, and normalization, while hyperparameters were optimized using Bayesian Optimization. The evaluation results show RMSE of 15.97 mm and MAE of 10.40 mm in Pekanbaru, indicating that the model achieves reasonably good accuracy. The 2025 prediction estimates the onset of the rainy season in January (first dekad) and the onset of the dry season in July (first dekad), consistent with BMKG forecasts. Recommended planting schedules for chili are January–February for the first planting season and November–December for the second. External validation in Malang, East Java, produced RMSE of 11.76 mm and MAE of 6.01 mm, suggesting that the model performs better in regions with distinct monsoonal rainfall patterns. These findings highlight the applicability of CNN–BiLSTM as a data-driven approach to support climate-based agricultural planning.
Keywords : CNN–BiLSTM, red chili, Pekanbaru, rainfall prediction, planting time

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2107135424

Penerbit

Pekanbaru : Universitas Riau – F.TEKNIK – INFORMATIKA., 2025

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2107135424

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

TEKNIK INFORMATIKA

Info Detail Spesifik

-

Pernyataan Tanggungjawab

daus

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar