時間序列資料的大致走向通常稱之為「趨勢線」，然而趨勢線未有精準描述的定義，每個人心中對趨勢線的形狀認知有些許差異，難以用一種趨勢線滿足所有人。另外個別使用者可能也不容易清楚敘述其心中的趨勢線樣貌。
本論文提出一個框架讓個別使用者以「手繪」的方式在十張時間序列資料上標出他認定的趨勢線，讓機器學習模型從中學習該使用者心中的趨勢線樣貌，以應用在其他時間序列資料上。

The tendency of a time series is usually referred to as a “trend line”. However, the precise definition of a trend line is still ambiguous. Given a time series, different users may come up with varying shapes of trend lines – some may prefer smooth lines, while others may hope the trend line responds to local turbulence. Therefore, a single trend line definition is challenging to meet everyone’s needs. Meanwhile, it could be complicated for users to clearly describe the requirements of a trend line in their minds.
This thesis proposes a framework to learn the customized trend lines that meet users’ demands. First, the framework asks users to plot the expected trend lines on ten time-series datasets. The framework then learns users’ preferred shapes and automatically draws the customized trend lines for other time-series datasets.

[1] H. Musbah, H. H. Aly, and T. A. Little, “A novel approach for seasonality and trend detection using fast fourier transform in box-jenkins algorithm,” in 2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE), 2020, pp. 1–5. doi: 10.1109/CCECE47787.2020.9255819.
[2] Y. R and S. Swamy, “Recent trends in time series forecasting-a survey,” vol. 7, pp. 5623–5628, Apr. 2020.
[3] J. Verbesselt, R. Hyndman, G. Newnham, and D. Culvenor, “Detecting trend and seasonal changes in satellite image time series,” Remote Sensing of Environment, vol. 114, no. 1, pp. 106–115, 2010, issn: 0034-4257. doi: https://doi.org/10. 1016/j.rse.2009.08.014. [Online]. Available: https://www.sciencedirect. com/science/article/pii/S003442570900265X.
[4] S.-J. Kim, K. Koh, S. Boyd, and D. Gorinevsky, “Ell_1 trend filtering,” SIAM review, vol. 51, no. 2, pp. 339–360, 2009.
[5] R. Hodrick and E. Prescott, “Postwar u.s. business cycles: An empirical investi- gation,” Journal of Money, Credit and Banking, vol. 29, no. 1, pp. 1–16, 1997. [Online]. Available: https://EconPapers.repec.org/RePEc:mcb:jmoncb:v:29: y:1997:i:1:p:1-16.
[6] R. B. Cleveland, W. S. Cleveland, J. E. McRae, and I. Terpenning, “Stl: A seasonal- trend decomposition,” J. Off. Stat, vol. 6, no. 1, pp. 3–73, 1990.
[7] Q. Wen, J. Gao, X. Song, L. Sun, H. Xu, and S. Zhu, “Robuststl: A robust seasonal- trend decomposition algorithm for long time series,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 33, 2019, pp. 5409–5416.
[8] J. L. Ticknor, “A bayesian regularized artificial neural network for stock market forecasting,” Expert systems with applications, vol. 40, no. 14, pp. 5501–5506, 2013.
[9] K.-j. Kim, “Financial time series forecasting using support vector machines,” Neu- rocomputing, vol. 55, no. 1-2, pp. 307–319, 2003.
[10] S. Babar and R. H, “Analysis of south west monsoon rainfall trend using statistical techniques over nethravathi basin,” International Journal of Advanced Technology in Civil Engineering, 2013.
49
參考文獻
[11] A. Baheti and D. Toshniwal, “Trend analysis of time series data using data min- ing techniques,” in 2014 IEEE International Congress on Big Data, IEEE, 2014, pp. 430–437.
[12] G. Mahalakshmi, S. Sridevi, and S. Rajaram, “A survey on forecasting of time series data,” in 2016 International Conference on Computing Technologies and Intelligent Data Engineering (ICCTIDE’16), IEEE, 2016, pp. 1–8.
[13] X. Yan and N. Chowdhury, “A comparison between svm and lssvm in mid-term electricity market clearing price forecasting,” May 2013, pp. 1–4, isbn: 978-1-4799- 0031-2. doi: 10.1109/CCECE.2013.6567685.
[14] N. Laptev, J. Yosinski, L. E. Li, and S. Smyl, “Time-series extreme event forecasting with neural networks at uber,” in International conference on machine learning, sn, vol. 34, 2017, pp. 1–5.
[15] G. P. Zhang and M. Qi, “Neural network forecasting for seasonal and trend time series,” European journal of operational research, vol. 160, no. 2, pp. 501–514, 2005.
[16] J. Park, C. G. Park, and K. Lee, “Simultaneous outlier detection and variable selec- tion via difference-based regression model and stochastic search variable selection,” Communications for Statistical Applications and Methods, vol. 26, pp. 149–161, Mar. 2019. doi: 10.29220/CSAM.2019.26.2.149.
[17] I. Choi, C. G. Park, and K. Lee, “Outlier detection and variable selection via difference based regression model and penalized regression,” Journal of the Korean Data And Information Science Sociaty, vol. 29, pp. 815–825, May 2018. doi: 10. 7465/jkdi.2018.29.3.815.
[18] Z. Liu and M. Hauskrecht, “Learning adaptive forecasting models from irregularly sampled multivariate clinical data,” in Thirtieth AAAI conference on artificial in- telligence, 2016.
[19] Q. Zeng, D. Li, G. Huang, et al., “Time series analysis of temporal trends in the pertussis incidence in mainland china from 2005 to 2016,” Scientific reports, vol. 6, no. 1, pp. 1–8, 2016.
[20] L. Anghinoni, L. Zhao, D. Ji, and H. Pan, “Time series trend detection and forecast- ing using complex network topology analysis,” Neural Networks, vol. 117, pp. 295– 306, 2019.
[21] A. De Livera, R. Hyndman, and R. Snyder, “Forecasting time series with complex seasonal patterns using exponential smoothing,” Journal of the American Statis- tical Association, vol. 106, pp. 1513–1527, Jan. 2010. doi: 10.1198/jasa.2011. tm09771.
[22] C. Sax and D. Eddelbuettel, “Seasonal adjustment by X-13ARIMA-SEATS in R,” Journal of Statistical Software, vol. 87, no. 11, pp. 1–17, 2018. doi: 10.18637/jss. v087.i11.
50

參考文獻
[23] D. F. Findley, B. C. Monsell, W. R. Bell, M. C. Otto, and B.-C. Chen, “New capabil- ities and methods of the x-12-arima seasonal-adjustment program,” Journal of Busi- ness Economic Statistics, vol. 16, no. 2, pp. 127–152, 1998, issn: 07350015. [Online]. Available: http://www.jstor.org/stable/1392565 (visited on 06/12/2022).
[24] A. Dokumentov and R. J. Hyndman, “Str: Seasonal-trend decomposition using regression,” INFORMS Journal on Data Science, 2021.
51