請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/98746完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 王勝德 | zh_TW |
| dc.contributor.advisor | Sheng-De Wang | en |
| dc.contributor.author | 戴廷磬 | zh_TW |
| dc.contributor.author | Ting-Ching Tai | en |
| dc.date.accessioned | 2025-08-18T16:19:43Z | - |
| dc.date.available | 2025-08-19 | - |
| dc.date.copyright | 2025-08-18 | - |
| dc.date.issued | 2025 | - |
| dc.date.submitted | 2025-08-07 | - |
| dc.identifier.citation | [1] Y.Liuet al.,“Itransformer: Inverted transformers are effective for time series forecasting,” arXiv preprint arXiv:2310.06625, 2023.
[2] Y. Nie, N. H. Nguyen, P. Sinthong, and J. Kalagnanam, “A time series is worth 64 words:Long-term forecasting with transformers,”arXivpreprintarXiv:2211.14730,2022. [3] Y. Zhang and J. Yan, “Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting,” in The eleventh international conference on learning representations, 2023. [4] T. Zhou, Z. Ma, Q. Wen, X. Wang, L. Sun, and R. Jin, “Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting,” in International conference on machine learning, PMLR, 2022, pp. 27268–27286. [5] Y.Liu, H. Wu,J. Wang, and M.Long, “Non-stationary transformers: Exploring the stationarity in time series forecasting,” Advances in neural information processing systems, vol. 35, pp. 9881–9893, 2022. [6] V. Ashish, “Attention is all you need,” Advances in neural information processing systems, vol. 30, p. I, 2017. [7] C. Challu, K. G. Olivares, B. N. Oreshkin, F. G. Ramirez, M. M. Canseco, and A. Dubrawski, “Nhits: Neural hierarchical interpolation for time series forecasting,” in Proceedings of the AAAI conference on artificial intelligence, vol. 37, 2023,pp. 6989–6997. [8] A.Das, W.Kong,A.Leach, S. Mathur, R. Sen, and R. Yu, “Long-term forecasting with tide: Time-series dense encoder,” arXiv preprint arXiv:2304.08424, 2023. [9] P. Tang and W. Zhang, “Pdmlp: Patch-based decomposed mlp for long-term time series forecasting,” arXiv preprint arXiv:2405.13575, 2024. [10] S. Wang et al., “Timemixer: Decomposable multiscale mixing for time series forecasting,” arXiv preprint arXiv:2405.14616, 2024. [11] S.-A. Chen, C.-L. Li, N. Yoder, S. O. Arik, and T. Pfister, “Tsmixer: An all-mlp architecture for time series forecasting,” arXiv preprint arXiv:2303.06053, 2023. [12] G.Woo,C.Liu,D.Sahoo,A.Kumar,andS.Hoi,“Etsformer:Exponential smoothing transformers for time-series forecasting,”arXiv preprint arXiv:2202.01381,2022. [13] B.N.Oreshkin,D.Carpov,N.Chapados,andY.Bengio,“N-beats:Neural basis expansion analysis for interpretable time series forecasting,”arXiv preprint arXiv:1905.10437,2019. [14] D. Salinas, V. Flunkert, J. Gasthaus, and T. Januschowski, “Deepar: Probabilistic forecasting with autoregressive recurrent networks,” International journal of forecasting, vol. 36, no. 3, pp. 1181–1191, 2020. [15] J. G. Zilly, R. K. Srivastava, J. Koutnık, and J. Schmidhuber, “Recurrent highway networks,” in International conference on machine learning, PMLR, 2017,pp. 4189–4198. [16] G. Lai, W.-C. Chang, Y. Yang, and H. Liu, “Modeling long-and short-term temporal patterns with deep neural networks,” in The 41st international ACM SIGIR conference on research & development in information retrieval, 2018, pp. 95–104. [17] I.O.Tolstikhin et al., “Mlp-mixer: An all-mlp architecture for vision,” Advances in neural information processing systems, vol. 34, pp. 24261–24272, 2021. [18] A. Dosovitskiy et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929, 2020. [19] Z.Gong,Y.Tang,andJ.Liang,“Patchmixer:A patch-mixing architecture for long term time series forecasting,” arXiv preprint arXiv:2310.00655, 2023. [20] K.Yietal.,“Frequency-domain mlps are more effective learners in time series forecasting,” Advances in Neural Information Processing Systems, vol. 36, pp. 7665676679, 2023. [21] Y.Zhang, X. Zhou, Y. Zhang, S. Li, and S. Liu, “Improving time series forecasting in frequency domain using a multi resolution dual branch mixer with noise insensitive arctanloss,” Scientific Reports, vol. 15, no. 1, p. 12557, 2025. [22] H. Zhou et al., “Informer: Beyond efficient transformer for long sequence time series forecasting,” in Proceedings of the AAAI conference on artificial intelligence,vol. 35, 2021, pp. 11106–11115. [23] H. Wu, J. Xu, J. Wang, and M. Long, “Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting,” Advances in neural information processing systems, vol. 34, pp. 22419–22430, 2021. [24] H. Wu, T. Hu, Y. Liu, H. Zhou, J. Wang, and M. Long, “Timesnet: Temporal 2d variation modeling for general time series analysis,”arXiv preprint arXiv:2210.02186,2022. [25] M.Liu et al., “Scinet: Time series modeling and forecasting with sample convolution and interaction,” Advances in Neural Information Processing Systems, vol. 35,pp. 5816–5828, 2022. [26] M.M.N.Murad,M.Aktukmak,andY.Yilmaz,“Wpmixer:Efficient multi-resolution mixing for long-term time series forecasting,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 39, 2025, pp. 19581–19588. [27] Z. Li et al., “Fourier neural operator for parametric partial differential equations,”arXiv preprint arXiv:2010.08895, 2020. [28] M. Li and Z. Zhu, “Spatial-temporal fusion graph neural networks for traffic flow forecasting,” in Proceedings of the AAAI conference on artificial intelligence,vol.35,2021, pp. 4189–4196. [29] J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 7132-7141. [30] Y. Hu, P. Liu, P. Zhu, D. Cheng, and T. Dai, “Adaptive multi-scale decomposition framework for time series forecasting,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 39, 2025, pp. 17359–17367. [31] P. Tang and W. Zhang, “Unlocking the power of patch: Patch-based mlp for long term time series forecasting,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 39, 2025, pp. 12640–12648. [32] Z. Li, S. Qi, Y. Li, and Z. Xu, “Revisiting long-term time series forecasting: An investigation on linear mapping,” arXiv preprint arXiv:2305.10721, 2023. [33] A. Zeng, M. Chen, L. Zhang, and Q. Xu, “Are transformers effective for time series forecasting?” In Proceedings of the AAAI conference on artificial intelligence,vol. 37, 2023, pp. 11121–11128. [34] L. Han, X.-Y. Chen, H.-J. Ye, and D.-C. Zhan, “Softs: Efficient multivariate time series forecasting with series-core fusion,” Advances in Neural Information Processing Systems, vol. 37, pp. 64145–64175, 2024. [35] D. Campos, M. Zhang, B. Yang, T. Kieu, C. Guo, and C. S. Jensen, “Lightts:Lightweight time series classification with adaptive ensemble distillation,” Proceedings of the ACM on Management of Data, vol. 1, no. 2, pp. 1–27, 2023. [36] S.Liuetal., “Pyraformer: Low-complexity pyramidal attention for long-range time series modeling and forecasting,”in#PLACEHOLDER_PARENT_METADATA_VALUE#,2022. [37] S. Li et al., “Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting,” Advances in neural information processing systems, vol. 32, 2019. [38] N. Kitaev, Ł. Kaiser, and A. Levskaya, “Reformer: The efficient transformer,”arXiv preprint arXiv:2001.04451, 2020. [39] Z. Li, Z. Rao, L. Pan, and Z. Xu, “Mts-mixers: Multivariate time series forecasting via factorized temporal and channel mixing,” arXiv preprint arXiv:2302.04501,2023. | - |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/98746 | - |
| dc.description.abstract | 準確預測多變量時間序列對於工業應用至關重要，然而，由於數據中固有的短期瞬態變化、長期相依性、隱藏的週期性，以及嚴格的計算效率要求，這項任務充滿挑戰。為了解決這些問題，我們提出了AMTF-MLP，一個創新的純多層感知器(MLP)架構，它透過自適應融合機制，整合了多尺度時域混合器與頻域頻譜學習。AMTF-MLP透過並行分支處理信號：一個帶有層級化區塊混合器的時域分支，用以捕捉局部與全域的時間模式；以及一個帶有頻譜MLP的頻域分支，用以建模週期性。
在多樣化的公開基準上進行的大量實驗，驗證了我們模型的有效性與通用性。在長期預測方面，AMTF-MLP展現了高度的競爭力，與iTransformer和PatchTST等主流Transformer模型相比，其均方誤差(MSE)降低了9.3%至20.4%。在高頻率的短期預測場景中，它取得了優異的成果，在PEMS數據集上，其MSE分別比AMD和iTransformer等強力競爭對手降低了24.3%和40.4%。 此模型優異的跨域性能，同時也體現在計算效率上。在其高效能的MLP同類模型中，AMTF-MLP展現了優異的記憶體用量，記憶體消耗比AMD少1.80倍，同時維持著快1.5倍的訓練速度。消融實驗證實，我們設計的每個組件都對模型 的效能至關重要。憑藉其線性複雜度與強大的實證結果，AMTF-MLP為真實世界的預測系統提供了一個強大且實用的解決方案。 | zh_TW |
| dc.description.abstract | Accurately predicting multivariate time series is essential for industrial applications, yet it poses significant challenges due to short-term transients, long-term dependencies, and hidden periodicities, alongside stringent computational efficiency requirements.To address these issues, we introduce AMTF-MLP, an innovative pure Multi-Layer Perceptron (MLP) architecture that integrates multi-scale time-domain mixers and frequency domain spectral learning, unified through adaptive fusion. AMTF-MLP processes the signal in parallel branches: a time-domain branch with hierarchical patch mixers to capture local and global temporal patterns, and a frequency-domain branch with a spectral MLP to model periodicities.
Extensive experiments on diverse public benchmarks validate our model's effectiveness and versatility. In long-term forecasting, it delivers highly competitive performance, reducing Mean Squared Error (MSE) by 9.3% to 20.4% compared to prominent Transformer models like iTransformer and PatchTST. In high-frequency short-term scenarios, it achieves leading results, reducing MSE by 24.3% and 40.4% against strong competitors like AMD and iTransformer, respectively, on the PEMS datasets. The model's excellent cross-domain performance is also reflected in its computational efficiency: among its high-performance MLP peers, it exhibits the most superior memory efficiency, consuming 1.8 × less memory than AMD, while maintaining a training speed 1.5 × faster. Ablation studies confirm that each component of our design is critical to the model’s efficacy. With its linear complexity and strong empirical results, AMTF-MLP presents a powerful and practical solution for real-world forecasting systems. | en |
| dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2025-08-18T16:19:43Z No. of bitstreams: 0 | en |
| dc.description.provenance | Made available in DSpace on 2025-08-18T16:19:43Z (GMT). No. of bitstreams: 0 | en |
| dc.description.tableofcontents | Verification Letter from the Oral Examination Committee i
誌謝 iii 摘要 iv Abstract v Contents vii List of Figures ix List of Tables xi Chapter 1 Introduction 1 1.1 Background and Motivation 1 1.2 Our Perspective 3 1.3 Proposed Approach—AMTF MLP 4 1.4 Contributions 5 Chapter 2 Related Works 7 2.1 Full-MLP Architectures and Model Simplicity 7 2.2 Patching Mechanism and Multi-Scale Time Series Modeling 9 2.3 Frequency-Domain Modeling and Noise Suppression 11 2.4 Synthesis and Research Gap 14 Chapter 3 Approach 17 3.1 Problem Definition 18 3.2 Input Embedding Layer 19 3.2.1 Input Embedding Layer 19 3.2.2 Time-Domain Embedding 21 3.2.3 Frequency-Domain Embedding 22 3.3 Multi-Scale Time-Domain Branch 22 3.3.1 Local Patch Mixer 24 3.3.2 Global Patch Mixer 25 3.3.3 Multi-Scale Feature Fusion 26 3.4 Frequency-Domain Branch 27 3.4.1 Frequency Feature Extraction(FFT+Spectral MLP) 28 3.4.2 Time Signal Reconstruction(IFFT & Time Signal Reconstruction) 29 3.4.3 Frequency Branch Output 30 3.5 Adaptive Fusion Module 30 Chapter 4 Experiments 33 4.1 Datasets and Implementation Details 33 4.2 Long-term Forecasting Results 35 4.3 Short-term Forecasting Results 37 4.4 Imputation Task Results 38 4.5 Model Efficiency During Training 40 4.6 Memory Usage During Training 41 Chapter 5 Ablation Study 43 5.1 Effect of the Multi-Scale Time-Domain Branch 43 5.2 Effect of the Frequency-Domain Branch 44 5.3 Effect of the Adaptive Fusion Mechanism 45 5.4 Ablation Study on Short-Term Forecasting 47 Chapter 6 Conclusion 49 References 51 | - |
| dc.language.iso | en | - |
| dc.subject | 多變量時間序列預測 | zh_TW |
| dc.subject | 純多層感知器模型 | zh_TW |
| dc.subject | 頻域分析 | zh_TW |
| dc.subject | 自適應特徵融合 | zh_TW |
| dc.subject | 多尺度時域建模 | zh_TW |
| dc.subject | 長期預測 | zh_TW |
| dc.subject | 短期預測 | zh_TW |
| dc.subject | MLP-based Models | en |
| dc.subject | Short-Term Forecasting | en |
| dc.subject | Long-Term Forecasting | en |
| dc.subject | Multi-Scale Temporal Modeling | en |
| dc.subject | Adaptive Feature Fusion | en |
| dc.subject | Frequency-Domain Analysis | en |
| dc.subject | Multivariate Time Series Forecasting | en |
| dc.title | 基於適應性多尺度時頻域多層感知器於時間序列預測之架構設計 | zh_TW |
| dc.title | AMTF-MLP: Adaptive Multi-Scale Time-Frequency MLP for Time Series Forecasting | en |
| dc.type | Thesis | - |
| dc.date.schoolyear | 113-2 | - |
| dc.description.degree | 碩士 | - |
| dc.contributor.oralexamcommittee | 吳沛遠;陳永昇;雷欽隆 | zh_TW |
| dc.contributor.oralexamcommittee | Pei-Yuan Wu;Yeong-Sheng Chen;Chin-Laung Lei | en |
| dc.subject.keyword | 多變量時間序列預測,純多層感知器模型,頻域分析,自適應特徵融合,多尺度時域建模,長期預測,短期預測, | zh_TW |
| dc.subject.keyword | Multivariate Time Series Forecasting,MLP-based Models,Frequency-Domain Analysis,Adaptive Feature Fusion,Multi-Scale Temporal Modeling,Long-Term Forecasting,Short-Term Forecasting, | en |
| dc.relation.page | 54 | - |
| dc.identifier.doi | 10.6342/NTU202503526 | - |
| dc.rights.note | 同意授權(限校園內公開) | - |
| dc.date.accepted | 2025-08-12 | - |
| dc.contributor.author-college | 電機資訊學院 | - |
| dc.contributor.author-dept | 電機工程學系 | - |
| dc.date.embargo-lift | 2025-08-19 | - |
| 顯示於系所單位： | 電機工程學系 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-113-2.pdf 授權僅限NTU校內IP使用（校園外請利用VPN校外連線服務） | 4.92 MB | Adobe PDF |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。