PREDICTION OF CHICKEN MEAT PRICES IN EAST JAVA PROVINCE USING A LONG SHORT-TERM MEMORY ALGORITHM WITH PARTICLE SWARM OPTIMIZATION

Nisa', Choirun (2026) PREDICTION OF CHICKEN MEAT PRICES IN EAST JAVA PROVINCE USING A LONG SHORT-TERM MEMORY ALGORITHM WITH PARTICLE SWARM OPTIMIZATION. Undergraduate thesis, UPN Veteran Jawa timur.

	Text (cover) 22081010001.-cover.pdf Download (2MB)
	Text (bab 1) 22081010001.-bab1.pdf Download (435kB)
	Text (bab 2) 22081010001.-bab2.pdf Restricted to Repository staff only until 17 July 2029. Download (1MB)
	Text (bab 3) 22081010001.-bab3.pdf Restricted to Repository staff only until 17 July 2029. Download (1MB)
	Text (bab 4) 22081010001.-bab4.pdf Restricted to Repository staff only until 17 July 2029. Download (2MB)
	Text (bab 5) 22081010001.-bab5.pdf Download (421kB)
	Text (daftar pustaka) 22081010001.-daftarpustaka.pdf Download (357kB)
	Text (lampiran) 22081010001.-lampiran.pdf Restricted to Repository staff only until 17 July 2029. Download (373kB)

Abstract

Chicken meat represents one of the strategic food commodities with highly fluctuating prices, which can affect food stability and decision-making processes for business actors and government institutions. Therefore, an accurate forecasting method is required to model price movements effectively. This study proposes the implementation of the Long Short-Term Memory (LSTM) algorithm for forecasting broiler chicken meat prices in East Java Province and to analyze the effect of Particle Swarm Optimization (PSO) to improving the model's predictive performance through hyperparameter optimization. The dataset used in this study consists of daily meat prices in East Java from January 2018 to January 2026, obtained from the National Strategic Food Price Information Center (PIHPS) database. The research methodology comprises data preprocessing, feature engineering, data normalization using Min–Max Scaling, sequence generation with a 30-day window size, LSTM model development, hyperparameter optimization using PSO, and model evaluation using five-fold Walk-Forward Validation. Model performance was evaluated through Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results indicate that PSO-based hyperparameter optimization successfully improved the performance of the LSTM model. The best scenario was achieved using 15 particles and 10 iterations, resulting in an average MAE of 232.89, RMSE of 364.95, and MAPE of 0.6814%, while the baseline LSTM approach achieved an average MAE of 560.08, RMSE of 731.39, and MAPE of 1.6273%. The application of PSO improved performance by reducing MAE, RMSE, and MAPE by 58.42%, 50.10%, and 58.13%, respectively, compared to the baseline model. These results indicate that combining LSTM with PSO effectively enhances the accuracy of broiler chicken meat price forecasting in East Java Province.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Parlika, Rizky	NIDN0718058401	rizkyparlika.if@upnjatim.ac.id
Thesis advisor	Al Haromainy, Muhammad Muharrom	NIDN0701069503	muhammad.muharrom.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA76.87 Neural computers

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Choirun Nisa' nisa

Date Deposited:

17 Jul 2026 07:09

Last Modified:

17 Jul 2026 07:26

URI:

https://repository.upnjatim.ac.id/id/eprint/55810