智能電錶基礎設施 (SMI) 的大規模發展使客戶和電力公司之間能夠進行雙向通信。單個客戶可以監控負載評估他們的負載消耗。在高峰時段，過度的負載消耗是不可避免的。因此可能會向客戶收取多付的費用。為了避免過度的負荷消耗，需要發展短期負載預測。準確的負載預測使個別客戶避免過度耗能。我們建議進行每小時負荷預測，因為我們發現超出的耗能發生在不同的小時。然而，有幾個問題使每小時負荷預測具有挑戰性。動態行為會影響負載消耗模式，而且由於人為、技術和傳輸錯誤，負載數據可能會出錯和遺失。因此，我們提出了一種名為長短期記憶 (LSTM) 的深度學習模型，該模型具有數據準備策略。我們發現數據分析和特徵工程對於改善預測結果非常重要。我們將我們的結果與基於先前研究的其他方法進行了比較，我們獲得了比以前的研究更好的結果。

The massive development of Smart Meter Infrastructure (SMI) enables bi-directional communication between customers and the power company. The individual customer can monitor load evaluate their load consumption. During peak hours, excessive load consumption is unavoidable. Thus, overpayment might be charged to the customers. To avoid the excessive load consumption, short-term load forecasting need to develop. The accurate load prediction enables individual customers to avoid excessive load consumption. We propose to make an hourly load forecasting because we found that the exceeded load consumption occurs in different hours. However, several issues make hourly load forecasting challenging. The dynamic behavior affects the load consumption pattern, and also due to human, technical, and transmission errors, the load data might be wrong and lost. Thus, we propose a deep learning model named Long short-term memory (LSTM) with a data preparation strategy. We found that data analysis and feature engineering are pretty essential to improve the prediction result. We compared our results with other approaches based on previous studies. We obtained a better result than the previous studies.

References
[1] M. Alhussein, K. Aurangzeb, and S. I. Haider, “Hybrid cnnlstm
model for shortterm
individual
household load forecasting,” IEEE Access, vol. 8, pp. 180544–180557, October 2020.
[2] F. U. M. Ullah, A. Ullah, I. U. Haq, S. Rho, and S. W. Baik, “Shortterm
prediction of residential
power energy consumption via cnn and multilayer
bidirectional
lstm networks,” IEEE Access, vol. 8,
pp. 123369–123380, December 2020.
[3] W. Kong, Z. Y. Dong, Y. Jia, D. J. Hill, Y. Xu, and Y. Zhang, “Shortterm
residential load forecasting
based on lstm recurrent neural network,” IEEE Transactions on Smart Grid, vol. 10, pp. 841–851,
September 2019.
[4] N. Q. Uy and V. H. Nam, “A comparison of amqp and mqtt protocols for internet of things,” in 2019
6th NAFOSTED Conference on Information and Computer Science (NICS), pp. 292–297, March 2019.
[5] Y. Wang, Q. Chen, T. Hong, and C. Kang, “Review of smart meter data analytics: Applications,
methodologies, and challenges,” IEEE Transactions on Smart Grid, vol. 10, pp. 3125–3148, March
2019.
[6] P. Nair and I. Kashyap, “Hybrid preprocessing
technique for handling imbalanced data and detecting
outliers for knn classifier,” in 2019 International Conference on Machine Learning, Big Data, Cloud
and Parallel Computing (COMITCon), pp. 460–464, October 2019.
[7] R. K. Agrawal, F. Muchahary, and M. M. Tripathi, “Long term load forecasting with hourly predictions
based on longshorttermmemory
networks,” in 2018 IEEE Texas Power and Energy Conference
(TPEC), pp. 1–6, March 2018.
[8] K. Liang, F. Liu, and Y. Zhang, “Household power consumption prediction method based on selective
ensemble learning,” IEEE Access, vol. 8, pp. 95657–95666, May 2020.
[9] S. Chakraborty, J. Banik, S. Addhya, and D. Chatterjee, “Study of dependency on number of lstm units
for character based text generation models,” in 2020 International Conference on Computer Science,
Engineering and Applications (ICCSEA), pp. 1–5, July 2020.
[10] S. H. Rafi, NahidAlMasood,
S. R. Deeba, and E. Hossain, “A shortterm
load forecasting method
using integrated cnn and lstm network,” IEEE Access, vol. 9, pp. 32436–32448, February 2021.
[11] T. Zhang, T. Zhu, P. Xiong, H. Huo, Z. Tari, and W. Zhou, “Correlated differential privacy: Feature
selection in machine learning,” IEEE Transactions on Industrial Informatics, vol. 16, pp. 2115–2124,
August 2020.
[12] P. Kowalski and R. Smyk, “Review and comparison of smoothing algorithms for onedimensional
data
noise reduction,” in 2018 International Interdisciplinary PhD Workshop (IIPhDW), pp. 277–281, June
2018.
60
[13] W. Deng, Y. Guo, J. Liu, Y. Li, D. Liu, and L. Zhu, “A missing power data filling method based on
improved random forest algorithm,” Chinese Journal of Electrical Engineering, vol. 5, pp. 33–39,
December 2019.
[14] J. Viana, R. J. Bessa, and J. Sousa, “Load forecasting benchmark for smart meter data,” in 2019 IEEE
Milan PowerTech, pp. 1–6, August 2019.
[15] D. Long, N. Nassif, and A. S. Ours, “Prediction of energy consumption in buildings by system identification,”
in 2016 Future Technologies Conference (FTC), pp. 701–705, January 2016.
[16] Xinda Ke, Anjie Jiang, and Ning Lu, “Load profile analysis and shortterm
building load forecast for
a university campus,” in 2016 IEEE Power and Energy Society General Meeting (PESGM), pp. 1–5,
November 2016.
[17] N. Amjady, “Shortterm
hourly load forecasting using timeseries
modeling with peak load estimation
capability,” IEEE Transactions on Power Systems, vol. 16, pp. 498–505, August 2001.
[18] A. Papalexopoulos and T. Hesterberg, “A regressionbased
approach to shortterm
system load forecasting,”
IEEE Transactions on Power Systems, vol. 5, pp. 1535–1547, November 1990.
[19] H. Yiling and H. Shaofeng, “A shortterm
load forecasting model based on improved random forest
algorithm,” in 2020 7th International Forum on Electrical Engineering and Automation (IFEEA),
pp. 928–931, February 2020.
[20] M. Ali, Z. A. Khan, S. Mujeeb, S. Abbas, and N. Javaid, “Shortterm
electricity price and load forecasting
using enhanced support vector machine and knearest
neighbor,” in 2019 Sixth HCT Information
Technology Trends (ITT), pp. 79–83, April 2019.
[21] B. Sun, L. Ma, W. Cheng, W. Wen, P. Goswami, and G. Bai, “An improved knearest
neighbours
method for traffic time series imputation,” in 2017 Chinese Automation Congress (CAC), pp. 7346–
7351, January 2017.
[22] B. AlHelali,
Q. Chen, B. Xue, and M. Zhang, “Gpbased
feature selection and weighted knnbased
instance selection for symbolic regression with incomplete data,” in 2020 IEEE Symposium Series on
Computational Intelligence (SSCI), pp. 905–912, January 2020.
[23] S. Ryu, M. Kim, and H. Kim, “Denoising autoencoderbased
missing value imputation for smart
meters,” IEEE Access, vol. 8, pp. 40656–40666, February 2020.
[24] C. Guan, P. B. Luh, L. D. Michel, and Z. Chi, “Hybrid kalman filters for very shortterm
load forecasting
and prediction interval estimation,” IEEE Transactions on Power Systems, vol. 28, pp. 3806–3817,
June 2013.
[25] C. S. Chen, M. S. Kang, J. C. Hwang, and C. W. Huang, “Temperature effect to distribution system load
profiles and feeder losses,” IEEE Transactions on Power Systems, vol. 16, pp. 916–921, November
2001.
61
[26] T. Company, “2020 sustainability report,” Report, vol. 88, June 2020.
[27] A. Savitzky and M. J. E. Golay, “Smoothing and Differentiation of Data by Simplified Least Squares
Procedures.,” Anal. Chem., vol. 36, pp. 1627–1639, July 1964.
[28] C. Cai, Y. Tao, Q. Ren, and G. Hu, “Shortterm
load forecasting based on mblstm
neural network,”
in 2020 Chinese Automation Congress (CAC), pp. 5402–5406, January 2020.
[29] F. Chollet et al., “Keras.” https://github.com/fchollet/keras, 2015.
[30] L. lab, “Theano.” https://theano-pymc.readthedocs.io/en/latest/, 2018.
[31] K. A. Koparanov, K. K. Georgiev, and V. A. Shterev, “Lookback period, epochs and hidden states
effect on time series prediction using a lstm based neural network,” in 2020 28th National Conference
with International Participation (TELECOM), pp. 61–64, December 2020.