FORECASTING GOOGLE SEARCH TRENDS FOR THE KEYWORD “KOPI KENANGAN” USING SINGULAR SPECTRUM ANALYSIS (SSA)

Daniswara, Arya Fadhilah (2026) FORECASTING GOOGLE SEARCH TRENDS FOR THE KEYWORD “KOPI KENANGAN” USING SINGULAR SPECTRUM ANALYSIS (SSA). Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 19081010187_Cover.pdf Download (1MB)
	Text (Bab 1) 19081010187_Bab 1.pdf Download (164kB)
	Text (Bab 2) 19081010187_Bab 2.pdf Restricted to Repository staff only until 14 July 2028. Download (237kB)
	Text (Bab 3) 19081010187_Bab 3.pdf Restricted to Repository staff only until 14 July 2028. Download (185kB)
	Text (Bab 4) 19081010187_Bab 4.pdf Restricted to Repository staff only until 14 July 2028. Download (638kB)
	Text (Bab 5) 19081010187_Bab 5.pdf Download (155kB)
	Text (Daftar pustaka) 19081010187_Dapus.pdf Download (130kB)
	Text (Lampiran) 19081010187_Lampiran.pdf Restricted to Repository staff only until 14 July 2028. Download (307kB)

Abstract

This study aims to forecast the Google Search Trends index for the keyword “Kopi Kenangan” in Indonesia using the Singular Spectrum Analysis (SSA) method. The data used in this study were obtained from Google Trends for the period of January 2021 to June 2026 and were processed into 66 monthly observations. Google Trends data in this study are positioned as a proxy or early indicator of public search interest, not as sales data, transaction volume, number of consumers, or actual demand. The research stages are organized using the Knowledge Discovery in Databases (KDD) framework as the data-processing workflow, while the main forecasting method is SSA. The data were divided chronologically using 70% training data and 30% testing data, resulting in 46 training observations and 20 testing observations. Parameter testing was conducted using window length L = 6 to L = 23 and the number of reconstruction components r = 2 to min(L - 1, 12). To address validation limitations, this study adds naive, seasonal naive, moving average, and simple ARIMA-grid baselines, as well as one-step walk-forward validation. Model performance was evaluated using MAE, MSE, RMSE, MAPE, sMAPE, and MASE. The results show that the best SSA configuration was obtained at L = 7 and r = 5, with MAE of 16.65, MSE of 430.96, RMSE of 20.76, MAPE of 24.74%, sMAPE of 27.78%, and MASE of 2.3101. The 12-month forecasting results for July 2026 to June 2027 indicate that the search index for “Kopi Kenangan” is projected to range from 95.57 to 100.00. Based on these results, SSA can be applied to forecast the Google Search Trends index; however, the forecasting results should be interpreted carefully as public search interest tendency, not as sales or actual demand prediction.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Anggraeny, Fetty Tri	0711028201	fettyanggraeny.if@upnjatim.ac.id
Thesis advisor	Haromainy, Muhammad Muharrom Al	0701069503	muhammad.muharrom.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA76.6 Computer Programming

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Arya arya fadhilah

Date Deposited:

14 Jul 2026 08:17

Last Modified:

14 Jul 2026 08:17

URI:

https://repository.upnjatim.ac.id/id/eprint/55381