Peramalan Search Interest Lima Merek Baju Muslim Dan Gamis Di Indonesia Berdasarkan Data Google Trends Menggunakan Model Arimax Dan Jordan RNN

Maulida, Rizka (2019) Peramalan Search Interest Lima Merek Baju Muslim Dan Gamis Di Indonesia Berdasarkan Data Google Trends Menggunakan Model Arimax Dan Jordan RNN. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

Text TA_RIZKA MAULIDA_06211540000066.pdf - Accepted Version Restricted to Repository staff only until 1 October 2022. Download (4MB) | Request a copy

Abstract

Pola search interest tentang merek baju muslim dan gamis dapat dilihat pada Google Trends yang menggambarkan informasi tentang istilah yang paling terkenal. Tujuan dari penelitian ini adalah menerapkan metode-metode machine learning untuk meramalkan search interest merek-merek baju muslim dan gamis kemudian membandingkan hasilnya dengan metode-metode statistik. Penelitian ini berfokus pada dua metode machine learning, yaitu Feedforward Neural Network (FFNN) dan Jordan Recurrent Neural Network (JRNN) serta empat metode statistika, yaitu Naïve, Winters’ Exponential Smoothing, ARIMA, dan ARIMAX. Data bulanan search interest lima merek baju muslim dan gamis terkenal di Indonesia, yaitu Rabbani, Zoya, Dian Pelangi, Elzatta, dan Shafira digunakan sebagai studi kasus penelitian ini. Hasil menunjukkan terdapat perbedaan metode terbaik untuk merek yang berbeda. FFNN merupakan metode terbaik dalam meramalkan search interest empat merek baju muslim dan tiga merek gamis di Indonesia, JRNN merupakan metode terbaik untuk baju muslim Rabbani dan gamis Dian Pelangi, serta ARIMA merupakan metode terbaik untuk gamis Elzatta. Selain itu, sebagian besar hasil peramalan menunjukkan search interest tertinggi pada periode Bulan Mei dan Juni 2019.
================================================================================================
The pattern of the search interest about brand of Moslem clothes
and gamis can be seen at Google Trends which illustrate the information
about the most famous term. The objective of this research is to apply
machine learning methods for forecasting the trend of search interest
about brands of Moslem clothes and gamis and then compare the results
to statistical methods. This research focuses on two methods in machine
learning, i.e. Feedforward Neural Network (FFNN) and Jordan
Recurrent Neural Network (JRNN) and four statistical methods, i.e.
Naïve, Winters’ Exponential Smoothing, ARIMA, and ARIMAX. Monthly
data about search interest of five famous brands of Moslem clothes and
gamis in Indonesia, i.e. Rabbani, Zoya, Dian Pelangi, Elzatta, and
Shafira are used as case studies of this research. The results showed that
there were differences in the best methods for different brands. FFNN was
the best method for forecasting search interest of four brands of Moslem
clothes and three brands of gamis, JRNN was the best method for
forecasting search interest of Moslem clothes Rabbani and gamis Dian
Pelangi, and ARIMA was the best method for forecasting search interest
of gamis Elzatta. Moreover, most of forecasting results showed that the
highest search interest were on May and June 2019.

Item Type:	Thesis (Undergraduate)
Additional Information:	RSSt 519.53 Mau p-1 2019
Uncontrolled Keywords:	ARIMA, ARIMAX, baju muslim, Winters’ Exponential Smoothing, FFNN, gamis, Google Trends, Naïve, Peramalan, JRNN
Subjects:	H Social Sciences > HB Economic Theory > Economic forecasting--Mathematical models. Q Science > QA Mathematics > QA280 Box-Jenkins forecasting
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Rizka Maulida
Date Deposited:	29 Dec 2021 05:51
Last Modified:	29 Dec 2021 05:51
URI:	http://repository.its.ac.id/id/eprint/61889

Actions (login required)

View Item