Hybrid Deep Learning with Denoising Stacked Autoencoders for Agricultural Price Forecasting
125 / 132
Keywords:
Agricultural Price; Deep learning; Machine learning; Forecasting; Time series dataAbstract
The complex nature of agricultural price data, characterized by perishability, seasonality, and non-stationarity, often renders conventional forecasting
models inadequate. To address these challenges, this study proposes two hybrid learning frameworks: Stacked Autoencoder-based Deep Learning
(SAE-DL) and Denoising Stacked Autoencoder-based Deep Learning (DSAE-DL). These models represent a significant departure from existing
deep learning (DL) models, such as Multilayer Perceptron (MLP), Recurrent Neural Network (RNN), 1D Convolutional Neural Network (1D CNN),
Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU). Unlike these standard models, which process raw input directly and often
struggle with data volatility, our proposed hybrid approach integrates an autoencoder bottleneck to perform hierarchical feature extraction and
dimensionality reduction prior to forecasting. The SAE-DL model focuses on capturing the underlying compressed representation of the price series,
while the DSAE-DL introduces a stochastic denoising mechanism to specifically reconstruct core features from noisy data and outliers. Empirical
analysis using a real-world onion price dataset reveals that the proposed DSAE-DL models, particularly the DSAE-RNN configuration, consistently
outperform all other models across metrics, including RMSE, MAE, and MAPE. The Diebold-Mariano (DM) test further confirms that both proposed
hybrid frameworks, SAE-DL and DSAE-DL, significantly improve prediction accuracy compared to standalone DL models. These findings highlight
the superior robustness of hybrid autoencoder-based architectures in contending with market volatility.
Downloads
References
1. Avinash, G., V., R., Ray, M., Paul, R.K., Godara, S., Nayak, G.H.H.,
2. Kumar, R.R., Manjunatha, B., Dahiya, S., Iquebal, M.A. (2024).
3. Hidden Markov guided Deep Learning models for forecasting
4. highly volatile agricultural commodity prices. Applied Soft
5. Computing, 111557 https://doi.org/10.1016/j.asoc.2024. 111557
6. Chung, J., Gulcehre, C., Cho, K., Bengio, Y. (2014). Empirical
7. Evaluation of Gated Recurrent Neural Networks on Sequence
8. Modeling https://doi.org/10. 48550/ARXIV.1412.3555
9. Haykin, s. (2009). Neural Networks and Learning Machines, 3rd edn.
10. Pearson Education India.
11. Hinton, G.E., Salakhutdinov, R.R. (2006). Reducing the dimensionality
12. of data with neural networks. Science 313(5786), 504-507 https://
13. doi.org/10.1126/science. 1127647
14. Hochreiter, S., Schmidhuber, J. (1997). Long Short-Term Memory.
15. Neural Computation 9(8), 1735-1780 https://doi.org/10.1162/
16. neco.1997.9.8.1735
17. Liu, P., Zheng, P., Chen, Z. (2019). Deep Learning with Stacked
18. Denoising Auto-Encoder for Short-Term Electric Load
19. Forecasting. Energies 12(12), 2445 https: //doi.org/10.3390/
20. en12122445
21. Liu, Y., Zhao, C., Huang, Y. (2022). A Combined Model for Multivariate
22. Time Series Forecasting Based on MLP-Feedforward Attention
23. LSTM. IEEE Access 10, 88644-88654 https://doi.org/10.1109/
24. ACCESS.2022.3192430
25. Najafi, A., Homaee, O., Jasinski, M., Golshan, M., Leonowicz, Z.
26. (2022). Application of Extreme Learning Machine-Autoencoder
27. to Medium Term Electricity Price Forecasting. In: 2022 IEEE
28. International Conference on Environment and Electrical
29. Engineering and 2022 IEEE Industrial and Commercial Power
30. Systems Europe (EEEIC / I&CPS Europe), pp. 1-4. IEEE,
31. ikonomidis, A., Catal, C., Kassahun, A. (2023). Deep learning for
32. crop yield prediction: A systematic literature review. New Zealand
33. Journal of Crop and Horticultural Science 51(1), 1-26 https://doi.
34. org/10.1080/01140671.2022.2032213
35. Paul, R.K., Yeasin, Md., Kumar, P., Kumar, P., Balasubramanian,
36. M., Roy, H.S., Paul, A.K., Gupta, A. (2022). Machine learning
37. techniques for forecasting agricultural prices: A case of brinjal in
38. Odisha, India. PLOS ONE 17(7), 0270553 https://doi.org/10.1371/
39. journal.pone.0270553
40. Samal, K.K.R., Babu, K.S., Das, S.K. (2021). Temporal convolutional
41. denoising autoencoder network for air pollution prediction
42. with missing values. Urban Climate 38, 100872 https://doi.
43. org/10.1016/j.uclim.2021.100872
44. Soui, M., Smiti, S., Mkaouer, M.W., Ejbali, R. (2020). Bankruptcy
45. Prediction Using Stacked Auto-Encoders. Applied Artificial
46. Intelligence 34(1), 80-100 https://doi.org/10.1080/08839514.201
47. 9.1691849
48. Sun, F., Meng, X., Zhang, Y., Wang, Y., Jiang, H., Liu, P. (2023).
49. Agricultural Product 26 Price Forecasting Methods: A
50. Review. Agriculture 13(9), 1671 https:// doi.org/10.3390/
51. agriculture13091671
52. Vincent, P., Larochelle, H., Bengio, Y., Manzagol, P.-A. (2008).
53. Extracting and composing robust features with denoising
54. autoencoders. In: Proceedings of the 25th International Conference
55. on Machine Learning - ICML ’08, pp. 1096-1103. ACM Press,
56. Helsinki, Finland. https://doi.org/10.1145/1390156.1390294
57. Wang, L., Zhang, Z., Chen, J. (2017). Short-Term Electricity Price
58. Forecasting With Stacked Denoising Autoencoders. IEEE
59. Transactions on Power Systems 32(4), 2673-2681 https://doi.
60. org/10.1109/TPWRS.2016.2628873
61. Xiong, T., Li, C., Bao, Y. (2018). Seasonal forecasting of agricultural
62. commodity price using a hybrid STL and ELM method: Evidence
63. from the vegetable market in China. Neurocomputing 275,
64. 2831-2844 https://doi.org/10.1016/j. neucom.2017.11.053