The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

Prediction of electricity consumption in A hotel using ensemble learning with temperature

앙상블 학습과 온도 변수를 이용한 A 호텔의 전력소모량 예측

  • Kim, Jaehwi (Department of Statistics, Duksung Women's University) ;
  • Kim, Jaehee (Department of Statistics, Duksung Women's University)
  • 김재휘 (덕성여자대학교 정보통계학과) ;
  • 김재희 (덕성여자대학교 정보통계학과)
  • Received : 2019.01.18
  • Accepted : 2019.02.15
  • Published : 2019.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

Forecasting the electricity consumption through analyzing the past electricity consumption a advantageous for energy planing and policy. Machine learning is widely used as a method to predict electricity consumption. Among them, ensemble learning is a method to avoid the overfitting of models and reduce variance to improve prediction accuracy. However, ensemble learning applied to daily data shows the disadvantages of predicting a center value without showing a peak due to the characteristics of ensemble learning. In this study, we overcome the shortcomings of ensemble learning by considering the temperature trend. We compare nine models and propose a model using random forest with the linear trend of temperature.

과거의 전력소모량을 분석하여 미래의 전력소모량을 예측하는 것은 에너지 계획과 정책 결정에 있어 많은 이점을 가져다준다. 기계학습은 최근 전력소모량을 예측하는 분석 방법으로 많이 사용하고 있다. 그중 앙상블 학습은 모형의 과적합 현상을 방지하고 분산을 줄여 예측의 정확성을 높이는 방법으로 알려져 있다. 하지만 일별 데이터에 앙상블 학습을 적용했을 때 분석 방법의 특성으로 인해 피크를 잘 나타내지 못하고 중심값으로 예측하는 단점을 보였다. 본 연구에서는 앙상블 학습 전에 온도 변수와의 상관성을 고려하여 선형모형으로 적합함으로써 앙상블 학습의 단점을 보완한다. 그리고 9개의 모형을 비교한 결과 온도 변수를 선형모형으로 적합하고 랜덤포레스트를 사용한 모형이 결과가 가장 좋음을 보여준다.

Keywords

GCGHDE_2019_v32n2_319_f0001.png 이미지

Figure 2.1. Daily electricity consumption data of A hotel from 2016.1.1 to 2017.12.31.

GCGHDE_2019_v32n2_319_f0002.png 이미지

Figure 2.2. Comparison of weekly consumptions of A hotel during the first 16 weeks (2016.1–2016.4).

GCGHDE_2019_v32n2_319_f0003.png 이미지

Figure 2.3. Comparing with monthly consumptions of A hotel in 2016 and 2017.

GCGHDE_2019_v32n2_319_f0004.png 이미지

Figure 2.4. Daily average temperature and adjusted temperature in Suwon city.

GCGHDE_2019_v32n2_319_f0005.png 이미지

Figure 3.1. Regression tree example.

GCGHDE_2019_v32n2_319_f0006.png 이미지

Figure 4.1. Scatter plot between log-transformed electricity consumption and adjusted temperature.

GCGHDE_2019_v32n2_319_f0007.png 이미지

Figure 5.1. Comparison of random forest forecasts

Table 5.1. Parameters used in each model

GCGHDE_2019_v32n2_319_t0001.png 이미지

Table 5.2. Accuracy based on train and test data

GCGHDE_2019_v32n2_319_t0002.png 이미지

References

  1. Ahn, J. Y., Park, S. M., and Kim, C. B. (2017). A study on neural network model for winter electric power demand prediction, Journal of Korean Institute of Information Technology, 15, 1-9. https://doi.org/10.14801/jkiit.2017.15.12.1
  2. Choi, R. I. (2018). Tourism strategy and cultural policy research through development of local tourism resources, Asia-Pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology, 8, 43-51.
  3. Cleveland, R. B., Cleveland, W. S., McRae, J. E., and Terpenning, I. J. (1990). STL: A seasonal-trend decomposition procedure based on loess, Journal of Official Statistics, 6, 3-73.
  4. Cowpertwait, P. S. P. and Metcalfe, A. V. (2009). Introductory Time Series with R, Springer.
  5. Grmanova, G., Laurinec, P., Rozinajova, V., et al. (2016). Incremental ensemble learning for electricity load forecasting, Acta Polytechnica Hungarica, 13 97-117.
  6. Hothorn, T., Hornik, K., Strobl, C., and Zeileis, A. (2010). Party: A laboratory for recursive partytioning.
  7. James, G., Witten, D., Hastie, T., and Tibshirani, R. (2013). An Introduction to Statistical Learning with Applications in R, Springer, New York.
  8. Laurinec, P. (2017). Ensemble learning for time series forecasting in R, (2018, August 10), Retrieved from https://petolau.github.io/Ensemble-of-trees-for-forecasting-time-series/
  9. Lee, S. I. (2015). Transition of energy management paradigm and future demand management technology prospect, Future Horizon, 24, 12-15.
  10. Massidda, L. and Marrocu, M. (2018). Smart meter forecasting from one minute to one year horizons, Energies, 11 3520. https://doi.org/10.3390/en11123520
  11. Moritz, S. and Bartz-Beielstein, T. (2017). imputeTS: time series missing value imputation in R, The R Journal, 9 207-218. https://doi.org/10.32614/RJ-2017-009
  12. Park, J. H., Na, W. S., and Xu, Y. (2017). Evaluation of demand power prediction performance based on deep learning algorithm and data preprocessing. In Proceedings of the Korea Information Science Society, 1882-1884.
  13. Shi, H., Xu, M., and Li, R. (2018). Deep learning for household load forecasting-A novel pooling deep RNN, IEEE Transactions on Smart Grid, 9, 5271-5280. https://doi.org/10.1109/TSG.2017.2686012
  14. Shin, D. H. and Kim, C. B. (2016). A study on deep learning input pattern for summer power demand prediction, Journal of Korean Institute of Information Technology, 14, 127-134. https://doi.org/10.14801/jkiit.2016.14.11.127
  15. Tak, H., Kim, T., Cho, H. G., and Kim, H. (2016). A new prediction model for power consumption with local weather information, Journal of the Korea Contents Association, 16, 488-498.
  16. Therneau, T. M. and Atkinson, E. J. (1997). An introduction to recursive partitioning using the RPART routines.
  17. Yun, S. C. (2018). Focus on fair growth of contents industry and fostering high valued tourism industry, Nara Economy (nara.kdi.re.kr) 2018 Feb, Special Theme, 16-17.