OPTIMALISASI PREDIKSI PENJUALAN ROKOK MENGGUNAKAN PROPHET DENGAN REGRESSOR EKSTERNAL (STUDI KASUS: PT XYZ SEKTOR SULAWESI)

Ariq, Muhammad Akhtar (2026) OPTIMALISASI PREDIKSI PENJUALAN ROKOK MENGGUNAKAN PROPHET DENGAN REGRESSOR EKSTERNAL (STUDI KASUS: PT XYZ SEKTOR SULAWESI). Diploma thesis, UPN Veteran Jawa Timur.

This is the latest version of this item.

	Text (Cover) 22082010235_cover (2).pdf Download (803kB)
	Text (BAB 1) 22082010235_bab1.pdf Download (243kB)
	Text (BAB 2) 22082010235_bab2.pdf Restricted to Repository staff only until 17 July 2029. Download (580kB)
	Text (BAB 3) 22082010235_bab3.pdf Restricted to Repository staff only until 17 July 2029. Download (912kB)
	Text (BAB 4) 22082010235_bab4.pdf Restricted to Repository staff only until 17 July 2029. Download (604kB)
	Text (BAB 5) 22082010235_bab5.pdf Download (183kB)
	Text (Daftar Pustaka) 22082010235_daftarpustaka.pdf Download (190kB)
	Text (Lampiran) 22082010235_lampiran.pdf Restricted to Repository staff only until 17 July 2029. Download (1MB)

Abstract

Cigarette sales forecasting is needed to help companies plan distribution, manage inventory, and support data-based decision-making. This study aims to optimize the Prophet model for weekly cigarette sales forecasting at PT XYZ Sulawesi Sector by considering external factors that may influence sales changes. The data used in this study consist of weekly sales data from 2024 to 2025, measured in thousand sticks. The additional variables used include national holidays, Provincial Minimum Wage (UMP), Tobacco Excise/Retail Selling Price (CHT/HJE), sales lag, and an intervention dummy to indicate a period of sharp sales decline. The results show that the intervention-focused model was selected as the final model because it achieved the best validation result and was more suitable for the business context. In the testing stage, this model produced MAE of 0.5554, RMSE of 0.9086, MAPE of 116.8991%, SMAPE of 34.0708%, and WMAPE of 28.1474%. The high MAPE value was influenced by very small actual values during the sales shock period in August–September 2025. Based on company confirmation, this period was related to declining orders and overstock conditions. This study shows that the intervention dummy helps the model capture extreme pattern changes better than slow-moving macro variables

Item Type:

Thesis (Diploma)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Anindo, Saka Fitri	NIDN0025039303	anindo.saka.si@upnjatim.ac.id
Thesis advisor	Iqbal, Ramadhani Mukhlis	NIDN0705039304	Iqbal.ramadhani.fasilkom@upnjatim.ac.id

Subjects:

T Technology > T Technology (General) > T58.6-58.62 Management Information Systems

Divisions:

Faculty of Computer Science > Departemen of Information Systems

Depositing User:

Muhammad Akhtar Ariq

Date Deposited:

17 Jul 2026 09:00

Last Modified:

20 Jul 2026 02:01

URI:

https://repository.upnjatim.ac.id/id/eprint/56072