During the transportation of finished grain, particularly over long distances, dynamic fluctuations in temperature and humidity can affect the biochemical mechanisms of the grain, leading to increased energy loss and a higher risk of mold formation. However, existing prediction methods (e.g., LSTM, Informer) exhibit significant shortcomings in addressing these issues, especially in handling non-stationary time-series signals and cross-regional environmental coupling effects. To this end, this paper proposes a prediction method based on wavelet transform and the iTransformer model. This method innovatively constructs a dual-channel wavelet adaptive optimization module that dynamically selects the optimal wavelet basis functions for the temperature and humidity signals, respectively, to perform decomposition, achieving noise suppression and enhancement of key fluctuation features. Subsequently, the reconstructed features are input in parallel with the original data into the iTransformer model to improve prediction accuracy. Experimental results show that on the simulated Northeast-to-Yangtze-River-Delta transportation dataset, the proposed method achieves an MSE of 0.231 and an MAE of 0.233, which are significantly lower than those of the baseline and other prediction models. This research provides a new technical pathway for grain quality regulation in complex transportation environments.

Kun Wang, Fanghua Zhuo, Chao Yang, Muxiang Dai, Wenjing Fu, Lei Zhang, Bo Mao. Temperature and Humidity Prediction during Cross-Ecological Zone Transportation of Finished Grain Based on iTransformer and Wavelet Transform. Academic Journal of Computing & Information Science (2025), Vol. 8, Issue 9: 67-73. https://doi.org/10.25236/AJCIS.2025.080910.

[1] Ho S L, Xie M. The use of ARIMA models for reliability forecasting and analysis[J]. Computers & industrial engineering, 1998, 35(1-2): 213-216.

[2] Yu Y, Si X, Hu C, et al. A review of recurrent neural networks: LSTM cells and network architectures[J]. Neural computation, 2019, 31(7): 1235-1270.

[3] Dey R, Salem F M. Gate-variants of gated recurrent unit (GRU) neural networks[C]//2017 IEEE 60th international midwest symposium on circuits and systems (MWSCAS). IEEE, 2017: 1597-1600.

[4] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Advances in neural information processing systems, 2017, 30.

[5] Zhou H, Zhang S, Peng J, et al. Informer: Beyond efficient transformer for long sequence time-series forecasting[C]//Proceedings of the AAAI conference on artificial intelligence. 2021, 35(12): 11106-11115.

[6] Nie Y, Nguyen N H, Sinthong P, et al. A time series is worth 64 words: Long-term forecasting with transformers[J]. arXiv preprint arXiv:2211.14730, 2022.

[7] Liu Y, Hu T, Zhang H, et al. itransformer: Inverted transformers are effective for time series forecasting[J]. arXiv preprint arXiv:2310.06625, 2023.

[8] Zhang D. Wavelet transform[M]//Fundamentals of image data mining: Analysis, Features, Classification and Retrieval. Cham: Springer International Publishing, 2019: 35-44.

[9] Wu J, Wang Z. A hybrid model for water quality prediction based on an artificial neural network, wavelet transform, and long short-term memory[J]. Water, 2022, 14(4): 610.

[10] Zeng A, Chen M, Zhang L, et al. DLinear: Linear-based models for long-term time series forecasting[C]//Proceedings of the AAAI conference on artificial intelligence. 2023, 37(9): 11121-11128.

[11] Wang S, Wu H, Shi X, et al. Timemixer: Decomposable multiscale mixing for time series forecasting[J]. arXiv preprint arXiv:2405.14616, 2024.