時間序列分析能夠幫助我們找出資料中的時間趨勢和變化模式，透過對過去的觀測值進行分析，我們可以推斷未來的發展趨勢，並預測可能的模式和變化，提早做到決策與資源的分配。目前為止，已有許多相關研究提出不同的預測方法，然而真實的時序資料往往有其不穩定性及許多雜訊，因此，如何快速且精準地進行預測往往都是一大挑戰。單一的時間序列預測方法相對直觀而利於實現，但通常僅針對特定領域和特定問題進行設計，限制了應用範圍和可擴展性。有相關研究指出，若能結合訊號分解的前處理手法能有效地去除雜訊來捕捉到較穩定態樣，進而提升預測模型的準確度；然而，以訊號分解的處理手法會面臨到須適時更新每個子訊號的問題，若每當新資料產生時，便重新進行一次訊號分解的過程，則會花費大量時間在重複計算上面，因此，本研究將以增量學習的方式，對基於不同訊號分解方法中既有的子訊號進行更新，並結合長短期記憶模型觀察是否能在預測準確度與節省時間兩者間取得平衡；我們的實驗結果顯示，相較於小波分解，以自適應雜訊完備經驗模態分解的方法能取得較好的成效，而對於時間來說，由於電力資料適合的預測窗口更大，因此節省時間的效果則更為顯著。

Time series analysis can help us uncover trends and patterns in data over time. By analyzing past observations, we can infer future development trends and predict possible patterns and changes, enabling proactive decision-making and resource allocation. So far, numerous related studies have proposed various prediction methods. However, real-time series data often exhibit instability and noise, making rapid and accurate predictions a significant challenge. While individual time series prediction methods are relatively intuitive and practical, they are usually designed for specific domains and problems, limiting their applicability and scalability. Some research suggests that combining signal decomposition preprocessing techniques can effectively remove noise to capture more stable patterns, thereby improving the accuracy of prediction models. However, signal decomposition methods face the challenge of timely updating each sub-signal. If signal decomposition is performed every time new data is generated, it incurs a significant computational cost. In this study, we adopt an incremental learning approach to update existing sub-signals based on different signal decomposition methods. We integrate this approach with an LSTM prediction model to observe whether a balance can be achieved between prediction accuracy and time savings. Our experimental results show that, compared to wavelet decomposition, the CEEMDAN decomposition method yields better results. Additionally, for time considerations, as electricity data is suitable for a larger window size, the time-saving effects are more pronounced.

[1] P. Esling, and C. Agon. "Time-series data mining." ACM Computing Surveys (CSUR) 45.1 (2012): 1-34.
[2] J. -D. Hamilton. "Time series analysis." Princeton university press, 2020.
[3] R. -J. Hyndman, and G. Athanasopoulos. "Forecasting: principles and practice." OTexts, 2018.
[4] P. -C. Reiss, and M. -W. White. "Household electricity demand revisited." The Review of Economic Studies 72.3 (2005): 853-883.
[5] Z. Shao, et al. "A review of the decomposition methodology for extracting and identifying the fluctuation characteristics in electricity demand forecasting." Renewable and Sustainable Energy Reviews 75 (2017): 123-136.
[6] R. -S. Tsay. "Analysis of financial time series." John wiley & sons, 2005.
[7] W. -Y. Lin, Y. -H. Hu, and C. -F. Tsai. "Machine learning in financial crisis prediction: a survey." IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews) 42.4 (2011): 421-436.
[8] S. Siami-Namini, N. Tavakoli, and A. -S. Namin. "A comparison of ARIMA and LSTM in forecasting time series." 2018 17th IEEE international conference on machine learning and Applications (ICMLA). IEEE, 2018.
[9] E. -S. Gardner Jr. “Exponential smoothing: The state of the art.” Journal of Forecasting 4.1 (1985): 1-28.
[10] K. -H. Brodersen, et al. "Inferring causal impact using Bayesian structural time-series models." The Annals of Applied Statistics (2015): 247-274.
[11] M. -J. Kane, et al. "Comparison of ARIMA and Random Forest time series models for prediction of avian influenza H5N1 outbreaks." BMC Bioinformatics 15.1 (2014): 1-9.
[12] N. -I. Sapankevych, and R. Sankar. "Time series prediction using support vector machines: a survey." IEEE computational intelligence magazine 4.2 (2009): 24-38.
[13] I. -E. Livieris, E. Pintelas, and P. Pintelas. "A CNN–LSTM model for gold price time-series forecasting." Neural computing and applications 32 (2020): 17351-17360.
[14] S. Namini, N. Tavakoli, and A. -S. Namin. "The performance of LSTM and BiLSTM in forecasting time series." 2019 IEEE International Conference on Big Data (Big Data). IEEE, 2019.
[15] F. Karim, et al. "LSTM fully convolutional networks for time series classification." IEEE Access 6 (2017): 1662-1669.
[16] X. Chen, et al. "A study of time series noise reduction techniques in the context of land cover change detection." (2011).
[17] G. Chen, and Z. Wang. "A signal decomposition theorem with Hilbert transform and its application to narrowband time series with closely spaced frequency components." Mechanical systems and signal processing 28 (2012): 258-279.
[18] L. Zhi, L. Jian, and Cao. "Financial time series forecasting model based on CEEMDAN and LSTM." Physica A: Statistical mechanics and its applications 519 (2019): 127-139.
[19] X. An, et al. "Wind farm power prediction based on wavelet decomposition and chaotic time series." Expert Systems with Applications 38.9 (2011): 11280-11285.
[20] R. Faltermeier, et al. "Sliding empirical mode decomposition." The 2010 international joint conference on neural networks (IJCNN). IEEE, 2010.
[21] P. Stepien. "Sliding window empirical mode decomposition-its performance and quality." EPJ Nonlinear Biomedical Physics 2 (2014): 1-13.
[22] K. Yan, et al. "Multi-step short-term power consumption forecasting with a hybrid deep learning strategy." Energies 11.11 (2018): 3089.
[23] L. Chen, et al. "A hybrid attention-based EMD-LSTM model for financial time series prediction." 2019 2nd international conference on artificial intelligence and big data (ICAIBD). IEEE, 2019.
[24] J. Cao, Z. Li, and J. Li. "Financial time series forecasting model based on CEEMDAN and LSTM." Physica A: Statistical mechanics and its applications 519 (2019): 127-139.
[25] G. Rilling, F. Patrick, and G. Paulo. "On empirical mode decomposition and its algorithms." IEEE-EURASIP workshop on nonlinear signal and image processing. Vol. 3. No. 3. Grado: IEEE, 2003.
[26] Y. Gao, et al. "Analysis and solution to the mode mixing phenomenon in EMD." 2008 Congress on Image and Signal Processing. Vol. 5. IEEE, 2008.
[27] J. Wang, W. Liu, and S. Zhang. "An approach to eliminating end effects of EMD through mirror extension coupled with support vector machine method." Personal and Ubiquitous Computing 23 (2019): 443-452.
[28] D. Zhang, and D. Zhang. "Wavelet transform." Fundamentals of image data mining: Analysis, Features, Classification and Retrieval (2019): 35-44.
[29] P. Sugiartawan, R. Pulungan, and A. -K. Sari. "Prediction by a hybrid of wavelet transform and long-short-term-memory neural network." International Journal of Advanced Computer Science and Applications 8.2 (2017).
[30] L. Bello. "Machine learning for big sequence data: Wavelet-compressed Hidden Markov Models." (2020).
[31] M. Srivastava, C. -L. Anderson, and J. -H. Freed. "A new wavelet denoising method for selecting decomposition levels and noise thresholds." IEEE access 4 (2016): 3862-3877.
[32] D. Valencia, et al. "Comparison analysis between rigrsure, sqtwolog, heursure and minimaxi techniques using hard and soft thresholding methods." 2016 XXI Symposium on Signal Processing, Images and Artificial Vision (STSIVA). IEEE, 2016.
[33] H. Wang, M. Li, and X. Yue. "IncLSTM: incremental ensemble LSTM model towards time series data." Computers & Electrical Engineering 92 (2021): 107156.