Prediksi Harga Komoditas Bawang Merah di Kota Padang Menggunakan Algoritma K-Nearest Neighbors

Zakki, Zakki (2025) Prediksi Harga Komoditas Bawang Merah di Kota Padang Menggunakan Algoritma K-Nearest Neighbors. Skripsi thesis, UIN Imam Bonjol Padang.

	Text (Cover -Daftar Isi) Zakki.NIM 2117020078 Cover-Daftar isi.pdf - Published Version Download (508kB)
	Text (BAB I) Zakki.NIM 2117020078 BAB 1.pdf - Published Version Download (322kB)
	Text (BAB III) Zakki.NIM 2117020078 BAB 3.pdf - Published Version Download (418kB)
	Text (BAB V) Zakki.NIM 2117020078 BAB 5.pdf - Published Version Download (317kB)
	Text (Full Text) Zakki.NIM 2117020078 Full Text.pdf - Published Version Restricted to Repository staff only Download (4MB)

Abstract

Shallots are among Indonesia’s most important horticultural commodities, with prices that frequently fluctuate due to factors such as weather patterns, harvest volumes, and seasonal demand. This research seeks to forecast shallot prices in Padang City using the K-Nearest Neighbors (K-NN) algorithm within the CRISP-DM framework. The dataset comprises historical price records, rainfall, average temperature, average humidity, daily harvest volumes from Solok Regency, and major religious holidays, covering the period from August 1, 2024, to April 13, 2025. The methodology includes data cleaning, normalization, outlier detection using the IQR method, K-NN modeling, and performance evaluation through 5-Fold Cross Validation. The optimal configuration was achieved at K = 4 with distance weighting, producing an RMSE of 1698.7582 and an MAE of 1363.7646, which indicates that the model’s average prediction error is around Rp 1,364–1,699 per kg, and an R² of 0.8582. These results indicate strong predictive capability, as reflected in the high R² value and relatively low error rates. The model was deployed in a Streamlit-based application to provide accessible price forecasts for farmers, traders, and policymakers. This study is expected to contribute academically to the development of shallot price prediction models, while also serving as a practical foundation for more accurate and strategic decision-making in managing the shallot market dynamics in Padang City.

Actions (login required)

View Item