PENERAPAN MODEL HYBRID ARIMA-MLP UNTUK PERAMALAN JUMLAH PENUMPANG BUS TRANS JATIM

Moerrin, Naufal Baihaqi (2026) PENERAPAN MODEL HYBRID ARIMA-MLP UNTUK PERAMALAN JUMLAH PENUMPANG BUS TRANS JATIM. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21083010077_Cover_cropped-3-17.pdf Download (2MB)
	Text (Bab 1) 21083010077_Bab 1.pdf Download (321kB)
	Text (Bab 2) 21083010077_Bab 2.pdf Restricted to Repository staff only until 25 January 2029. Download (938kB)
	Text (Bab 3) 21083010077_Bab 3.pdf Restricted to Repository staff only until 25 January 2029. Download (608kB)
	Text (Bab 4) 21083010077_Bab 4.pdf Restricted to Repository staff only until 25 January 2029. Download (1MB)
	Text (Bab 5) 21083010077_Bab 5.pdf Download (262kB)
	Text (Daftar Pustaka) 21083010077_Daftar Pustaka.pdf Download (262kB)
	Text (Lampiran) 21083010077_Lampiran.pdf Restricted to Repository staff only until 25 January 2029. Download (446kB)

Abstract

Accurate passenger demand forecasting is essential to support operational planning and maintain service reliability in public transportation systems. Passenger demand time series commonly exhibit a combination of linear temporal structures and nonlinear fluctuations, making the use of a single forecasting model often less effective. This study employs a hybrid forecasting framework that integrates the Autoregressive Integrated Moving Average (ARIMA) model with a Multi-Layer Perceptron (MLP) neural network to predict short-term passenger demand. In this framework, ARIMA is used to model the linear components of the time series, such as trend, seasonality, and autocorrelation, following the Box– Jenkins procedure supported by stationarity testing and differencing. The residuals generated by the ARIMA model, which still contain nonlinear information, are subsequently modeled using an MLP trained through backpropagation to capture patterns that cannot be represented by linear models, so that the final forecasts are obtained by combining ARIMA outputs with nonlinear adjustments from the MLP. The analysis utilizes daily passenger count data of Bus Trans Jatim for the 2023– 2024 period, with preprocessing stages including exploratory time series analysis, variance stabilization, outlier detection, and differencing to ensure compliance with modeling assumptions. Forecasting performance is evaluated on test data using the Mean Absolute Percentage Error (MAPE). The results indicate that the hybrid ARIMA–MLP model achieves a MAPE of 4.95% and provides more adaptive shortterm forecasts than the standalone ARIMA model, with five-day-ahead forecasts showing demand patterns consistent with historical observations. In addition, the development of an interactive user interface facilitates model implementation and presentation of forecasting results, enabling predictions to be carried out more practically, quickly, and intuitively to support data-driven operational decisionmaking.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Damaliana, Aviolla Terza	NIDN0002089402	aviolla.terza.sada@upnjatim.ac.id
Thesis advisor	Diyasa, I Gede Susrama Mas	NIDN 0019067008	igsusrama.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA76.6 Computer Programming

Divisions:

Faculty of Computer Science > Departemen of Data Science

Depositing User:

Naufal Baihaqi Moerrin

Date Deposited:

30 Jan 2026 03:50

Last Modified:

30 Jan 2026 04:00

URI:

https://repository.upnjatim.ac.id/id/eprint/49147