Peramalan Harga Minyak Goreng di Kalimantan Timur Menggunakan Model Hybrid Time Series Regression Quadratic – Neural Network

Penulis

  • Risa Kristia Wahyuni Mulawarman Univeristy
  • Sri Wahyuningsih Program Studi Statistika, Jurusan Matematika, FMIPA Universitas Mulawarman
  • Meiliyani Siringoringo Program Studi Statistika, Jurusan Matematika, FMIPA Universitas Mulawarman

DOI:

https://doi.org/10.30872/eksponensial.v14i2.1123

Kata Kunci:

forecasting, hybrid, NN, TSR, Cooking oil price

Abstrak

A hybrid model is a combination of two or more forecasting methods. One of hybrid model that can be used in forecasting is Time Series Regression (TSR) Quadratic – Neural Network (NN). TSR Quadratic can be used in time series data that contains quadratic trend patterns, namely an increase or decrease that forms a curved or parabolic line NN is a method that has characteristics similar to biological neural networks in conducting data pattern recognition. This study was aimed to obtain a hybrid model of TSR quadratic-NN to forecast cooking oil prices in East Kalimantan and obtain forecasting results based on the best model. The results showed that the TSR Quadratic-NN hybrid model with 3 neurons in the hidden layer was the best model with a MAPE of 2.51368%. The forecasting results based on this model showed that cooking oil prices in East Kalimantan from January to December 2023 showed an increase

Unduhan

Diterbitkan

2023-11-30

Terbitan

Bagian

Articles