Claim Missing Document
Check
Articles

Found 1 Documents
Search

Performance Evaluation of Long Short-Term Memory for Chili Price Prediction Fikri, Fata Nabil; Nurochman, Nurochman
JISKA (Jurnal Informatika Sunan Kalijaga) Vol. 10 No. 1 (2025): January 2025
Publisher : UIN Sunan Kalijaga Yogyakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14421/jiska.2025.10.1.33-47

Abstract

Grocery prices often experience fluctuations in several regions of Indonesia, such as East Java Province. One of the commodities affected is chili, including both red chilies and bird’s eye chilies. Predictive steps that utilise machine learning, such as Long Short-Term Memory (LSTM), can be taken to estimate the next price of chili, with the expectation that the authorities can implement the appropriate strategy. LSTM is a network that was developed from RNN networks in previous times by offering a longer cell memory, allowing for the storage of more information. This research focuses on determining whether the LSTM network can be applied to the task of chili price prediction and identifying the suitable architecture and hyperparameter configuration for this case. For this reason, the experimental method is employed by testing several predetermined variables to determine the optimal architecture and hyperparameter configuration. The results of this research demonstrate that the LSTM network can be effectively applied in this case, and the obtained architecture and optimal hyperparameter configuration are consistent for both types of chilies, namely red chilies and bird’s eye chilies. For red chili, the best RMSE value that can be produced is 1751.890 and 1888.741 for bird’s eye chili.