DOI QR코드

DOI QR Code

Forecasting of Rental Demand for Public Bicycles Using a Deep Learning Model

딥러닝 모형을 활용한 공공자전거 대여량 예측에 관한 연구

  • Received : 2020.05.26
  • Accepted : 2020.06.17
  • Published : 2020.06.30

Abstract

This study developed a deep learning model that predicts rental demand for public bicycles. For this, public bicycle rental data, weather data, and subway usage data were collected. After building an exponential smoothing model, ARIMA model and LSTM-based deep learning model, forecasting errors were compared and evaluated using MSE and MAE evaluation indicators. Based on the analysis results, MSE 348.74 and MAE 14.15 were calculated using the exponential smoothing model. The ARIMA model produced MSE 170.10 and MAE 9.30 values. In addition, MSE 120.22 and MAE 6.76 values were calculated using the deep learning model. Compared to the value of the exponential smoothing model, the MSE of the ARIMA model decreased by 51% and the MAE by 34%. In addition, the MSE of the deep learning model decreased by 66% and the MAE by 52%, which was found to have the least error in the deep learning model. These results show that the prediction error in public bicycle rental demand forecasting can be greatly reduced by applying the deep learning model.

본 연구는 공공자전거의 대여량을 예측하는 딥러닝 모형을 개발하였다. 이를 위하여 공공자전거 대여량 자료, 기상 자료, 그리고 지하철 이용량 자료를 수집하였다. 지수평활 모형, ARIMA 모형과 LSTM기반의 딥러닝 모형을 구축한 후 MSE와 MAE 평가 지표를 사용하여 예측 오차를 비교·평가하였다. 평가 결과, 지수평활 모형으로 MSE 348.74, MAE 14.15 값이 산출되었다. ARIMA 모형으로 MSE 170.10, MAE 9.30 값을 얻었다. 그리고 딥러닝 모형으로 MSE 120.22, MAE 6.76 값이 산출되었다. 지수평활 모형의 값과 비교하여 ARIMA 모형의 MSE는 51%, MAE는 34% 감소하였다. 그리고 딥러닝 모형의 MSE는 66%, MAE는 52% 감소하여 딥러닝 모형의 오차가 가장 적은 것으로 파악되었다. 이러한 결과로부터 공공자전거 대여량 예측 분야에서 딥러닝 모형의 적용시 예측 오차를 크게 감소시킬 수 있을 것으로 판단된다.

Keywords

References

  1. Do M. and Noh Y.(2014), "Analysis of the Affecting Factors on the Bike-sharing Demand focused on Daejeon City," Journal of the Korean Society of Civil Engineers, vol. 34, no. 5, pp.1517-1524. https://doi.org/10.12652/Ksce.2014.34.5.1517
  2. Faghih-Imani A., Eluru N., El-Geneidy A. M., Rabbat M. and Haq U.(2014), "How land-use and urban form impact bicycle flows: Evidence from the bicycle-sharing system (BIXI) in Montreal," Journal of Transport Geography, vol. 41, pp.306-314. https://doi.org/10.1016/j.jtrangeo.2014.01.013
  3. Fu R., Zhang Z. and Li L.(2016), "Using LSTM and GRU Neural Network Methods for Traffic Flow Prediction," 31st Youth Academic Annual Conference of Chinese Association of Automation, p.329.
  4. Kaltenbrunner A., Meza R., Grivolla J., Codina J. and Banchs R.(2010), "Urban cycles and mobility patterns: Exploring and predicting trends in a bicycle-based public transport system," Pervasive and Mobile Computing, vol. 6, no. 4, pp.455-466. https://doi.org/10.1016/j.pmcj.2010.07.002
  5. Kim D., Shin H., Park J. and Im H.(2012), "The Impact of Weather on Bicycle Usage-Focus on Usage of Bike-sharing System in Goyang," Journal of Transport Research, vol. 19, no. 3, pp.77-88. https://doi.org/10.34143/jtr.2012.19.3.77
  6. Kim Y., Kim W., Park B., Park S., Park T., Oh H., Lee S., Lee Y., Lee J., Lim Y., Jeon J. and Cho S.(1989), Introduction to Statistics, Yongji-Munhwa, Korea.
  7. Lee J., Jeong G. and Shin H.(2016), "Impact Analysis of Weather Condition and Locational Characteristics on the Usage of Public Bike Sharing System," Journal of the Korean Society of Transportation, vol. 34, no. 5, pp.394-408. https://doi.org/10.7470/jkst.2016.34.5.394
  8. Lim H. and Chung K.(2019), "Development of Demand Forecasting Model for Seoul Shared Bicycle," Journal of the Korea Contents Association, vol. 19, no. 1, pp.132-140. https://doi.org/10.5392/JKCA.2019.19.01.132
  9. Min J., Mun H. and Lee Y.(2017), "Demand Forecast for Public Bicycles ("Tashu") in Daejeon using Random Forest," Proc. of the Korea Information Science Society Congress, p.969.
  10. Sa K.(2019), A Study on the Characteristics of Physical Environments Affecting the Usage of Public Bike in Seoul, Korea, Master's Thesis, Hanyang University.
  11. Yang H., Xie K., Ozbay K., Ma Y. and Wang Z.(2015), "Use of Deep Learning to Predict Daily Usage of Bike Sharing Systems," Transportation Research Record, vol. 2672, no. 36, pp.92-102. https://doi.org/10.1177/0361198118801354