提出了一种基于三维卷积和卷积长短期记忆（convolutional long short-term memory，CLSTM）神经网络的水产养殖溶解氧预测模型。首先，将输入向量及其转置相乘形成一个单通道矩阵，把一定时间段内的单通道矩阵堆叠成一个立方体作为输入数据；然后，将输入数据进行连续两次三维卷积来细化溶解氧相关因素的特征，并删除池化层以简化计算；最后，将三维卷积抽取的特征结果输入CLSTM模型以提取时间维度的信息，在全连接层根据梯度下降算法将数据反向更新。采集湖北省襄阳市某家特种水产养殖有限公司的实际数据进行实验。结果表明：相比于传统BP神经网络模型、Conv3D、Conv2D，所提出的模型具有更快的训练收敛速度、更高的预测精度和更好的预测稳定性，可以满足实际生产的需要。
In this paper, a neural network based on three-dimensional (3D) convolution and convolutional long short-term memory (CLSTM) is proposed to predict the dissolved oxygen in aquaculture. Firstly, an input vector is multiplied by its transpose to form a single-channel matrix, and the single-channel matrices within a certain period of time are stacked to form a cube as the input data. Secondly, two consecutive three-dimensional convolutions are carried out on the input data to refine the characteristics of dissolved oxygen related factors, and the pooling layer is deleted for reducing calculation. Finally,the feature results of 3D convolution extraction are sent to the CLSTM model for further information extraction of time dimension, and the data is updated reversely by the gradient descent algorithm through the full connection layer. The actual data of a special aquaculture company in Xiangyang, Hubei Province were collected for experiment, and experimental results show that the proposed model has faster training convergence speed, higher prediction accuracy and better prediction stability than traditional BP neural network models, Conv3D and Conv2D, and could meet the needs of actual production.
[1] Zhang X C, Endo M, Sakamoto T, et al. Studies on kuruma shrimp culture in recirculating aquaculture system with artificial ecosystem[J]. Aquaculture, 2018, 484:191-196.
[2] Li C, Li Z B, Wu J, et al. A hybrid model for dissolved oxygen prediction in aquaculture based on multi-scale features[J]. Information Processing in Agriculture, 2017, 1(5):11-20.
[3] Guseo R. Diffusion of innovations dynamics, biological growth and catenary function[J]. Physica A:Statistical Mechanics & Its Applications, 2016, 464:1-10.
[4] Scully E M. Mixing of dissolved oxygen in Chesapeake Bay driven by the interaction between wind-driven circulation and estuarine bathymetry[J]. Journal of Geophysical Research, 2016, 121(8):5639-5654.
[5] 王小艺, 赵晓平, 刘载文, 等. 基于灰色理论的湖库水体富营养化预测方法研究[J]. 计算机仿真, 2011, 28(1):17-19. Wang X Y, Zhao X P, Liu Z W, et al. Research on lake eutrophication forecasting methods based on grey theory[J]. Computer Simulation, 2011, 28(1):17-19. (in Chinese)
[6] Xiao Z, Peng L X, Chen Y, et al. The dissolved oxygen prediction method based on neural network[J]. Complexity, 2017:1-6.
[7] Liu S Y, Yan M X. Prediction of dissolved oxygen content in aquaculture of hyriopsis cumingii using Elman neural network[C]//2012 International Conference on Computer and Computing Technologies in Agriculture, 2012:508-518.
[8] Chen Y Y, Xu J, Yu H H, et al. Three-dimensional short-term prediction model of dissolved oxygen content based on PSO-BPANN algorithm coupled with Kriging interpolation[J]. Mathematical Problems in Engineering, 2016:6564202-6564210.
[9] Khotimah W N. Aquaculture water quality prediction using smooth SVM[J]. IPTEK Journal of Proceedings, 2014(1):342-345.
[10] Liu S Y, Tai H J, Ding Q S, et al. A hybrid approach of support vector regression with genetic algorithm optimization for aquaculture water quality prediction[J]. Mathematical & Computer Modelling, 2013, 58(3/4):458-465.
[11] Huan J, Cao W J, Qin Y L. Prediction of dissolved oxygen in aquaculture based on EEMD and LSSVM optimized by the Bayesian evidence framework[J]. Computers and Electronics in Agriculture, 2018, 150:257-265.
[12] Khadr M, Elshemy M. Data-driven modeling for water quality prediction case study:the drains system associated with Manzala Lake, Egypt[J]. Ain Shams Engineering Journal, 2017, 8(4):549-557.
[13] Sahoo M M, Patra K C, Khatua K K. Inference of water quality index using ANFIA and PCA[J]. Aquatic Procedia, 2015(4):1099-1106.
[14] Xu L Q, Liu S Y. Study of short-term water quality prediction model based on wavelet neural network[J]. Mathematical & Computer Modelling, 2013, 58(3/4):807-813.
[15] Shi P, Li G H, Yuan Y M, et al. Prediction of dissolved oxygen content in aquaculture using clustering-based Softplus extreme learning machine[J]. Computers and Electronics in Agriculture, 2019(157):329-338.
[16] Ta X X, Wei Y G. Research on a dissolved oxygen prediction method for recirculating aquaculture systems based on a convolution neural network[J]. Computers and Electronics in Agriculture, 2018(145):302-310.
[17] Ta X X, An D, Wei Y G. Dissolved oxygen prediction method for recirculating aquaculture system, based on a timing attenuation matrix and a convolutional neural network[J]. Aquaculture, 2019(503):26-33.
[18] Zeiler M D, Fergus R. Visualizing and understanding convolutional networks[C]//European Conference on Computer Vision, 2014:818-833.
[19] Jafrasteh B, Fathianpour N. A hybrid simultaneous perturbation artificial bee colony and back-propagation algorithm for training a local linear radial basis neural network on ore grade estimation[J]. Neurocomputing, 2017, 235:217-227.
[20] Schmidhuber J. Deep learning in neural networks:an overview[J]. Neural Networks, 2015, 61:85-117.
[21] Monir E C, Hicham O, Mohamed L. Scientific paper classification using convolutional neural networks[C]//International Conference Big Data and Internet Things, Tangier, Morocco, 2019:1-6.
[22] Yu R, Gao Y, Duan X, et al. QRS detection and measurement method of ECG paper based on convolutional neural networks[C]//201840th Annual International Conference of the IEEE Engineering in Medicine and Biology Society:4636-4639.
[23] Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural Computation, 1997, 8(9):1735-1780.
[24] Gers F A, Schmidhuber J. Recurrent nets that time and count[C]//Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks, 2000:189-194.
[25] Greff K, Srivastava R K, Koutník J, et al. LSTM:a search space odyssey[J]. IEEE Transactions on Neural Networks & Learning Systems, 2015, 28(10):2222-2232.
[26] Chung J, Gulcehre C, Cho K H, et al. Empirical evaluation of gated recurrent neural networks on sequence modeling[J]. Eprint Arxiv, 2014. 2014[2020-08-15]. https://arxiv.org/abs/1412.3555.
[27] Shi X J, Chen Z R, Wang H, et al. Convolutional LSTM network:a machine learning approach for precipitation nowcasting:NIPS'15[C]//Proceedings of the 28th International Conference on Neural Information Processing Systems, Cambridge, 2015:802-810.
