Sea surface height (SSH) prediction is theoretically and practically significant for global and regional ocean-related research. Numerous studies have been conducted to acquire accurate prediction results. However, most investigations on SSH ignore the importance of data at each time step on the prediction, which limits the accuracy of the final prediction. Therefore, a deep learning model combined Long Short-Term Memory (LSTM) network and Attention mechanism is proposed in this letter. This model integrates attention mechanism in both of time and space dimensions into LSTM. For time dimension, it assigns reasonable weight for data at each time step. For space dimension, it groups the data points close to each other, let model concentrate on points in the same group and eliminates the impact from other points. Daily absolute dynamic topography (ADT) in the South China Sea from January 2010 to December 2017 is adopted to conduct experiments. The proposed model demonstrates reliable results, the root mean square error is 0.38 cm, the mean absolute error is 0.0031, and the correlation coefficient reaches up to 0.999. The results show that the deep learning method based on attention mechanism is reliable for SSH prediction with high performance.

Ocean current (OC) prediction plays an important role for carrying out ocean-related activities. There are plenty of studies for OC prediction with deep learning to pursue better prediction performance, and the attention mechanism was widely used for these studies. However, the attention mechanism was usually combined with deep learning models rather than purely used to predict OC, or, if it was purely used, did not further optimize the attention weight. Therefore, a deep learning model based on weighted pure attention mechanism is proposed in this paper. This model uses the pure attention mechanism, introduces a weight parameter for the generated attention weight, and moves more attentions from other elements to the key elements based on weight parameter setting. To our knowledge, it is the first attempt to use the weighted pure attention mechanism to improve the OC prediction performance, and it is an innovation for OC prediction. The experiment results indicate that the proposed model can fully take advantage of the strengths from the pure attention mechanism; it can further optimize the pure attention mechanism and significantly improve the prediction performance, and is reliable for OC prediction with high performance for a wide time range and large spatial scope.

Multi-node wind speed forecasting is greatly important for offshore wind power. It is a challenging task due to unknown complex spatial dependencies. Recently, graph neural networks (GNN) have been applied to wind forecasting because of their capability in modeling dependencies. However, existing methods usually require a pre-defined graph structure, which is not optimal for the downstream task and limits the application scope of GNN. In this paper, we propose adaptive graph-learning convolutional networks (AGLCN) that can automatically infer hidden associations among multi-nodes through a graph-learning module. It simultaneously integrates the temporal and graph convolutional modules to capture temporal and spatial features in the data. Experiments are conducted on real-world multi-node wind speed data from the China Sea. The results show that our model achieves state-of-the-art results in all multi-scale wind speed predictions. Moreover, the learned graph can reveal spatial correlations from a data-driven perspective.

Applying prediction models to software engineering is an interesting research area. There have been many related studies which leverage prediction models to achieve good performance in various software engineering tasks. With more and more researches in software engineering leverage prediction models, there is a need to sort out related studies, aiming to summarize which software engineering tasks prediction models can apply to and how to better leverage prediction models in these tasks. This article conducts a comprehensive taxonomy on prediction models applied to software engineering. We review 136 papers from top conference proceedings and journals in the last decade and summarize 11 research topics prediction models can apply to. Based on the papers, we conclude several big challenges and directions. We believe that the comprehensive taxonomy will help us understand the research area deeper and infer several useful and practical implications.

Sea surface temperature (SST) prediction plays an important role in scientific research, environmental protection, and other marine-related fields. However, most of the current prediction methods are not effective enough to utilize the spatial correlation of SSTs, which limits the improvement of SST prediction accuracy. Therefore, this paper first explores spatial correlation mining methods, including regular boundary division, convolutional sliding translation, and clustering neural networks. Then, spatial correlation mining through a graph convolutional neural network (GCN) is proposed, which solves the problem of the dependency on regular Euclidian space and the lack of spatial correlation around the boundary of groups for the above three methods. Based on that, this paper combines the spatial advantages of the GCN and the temporal advantages of the long short-term memory network (LSTM) and proposes a spatiotemporal fusion model (GCN-LSTM) for SST prediction. The proposed model can capture SST features in both the spatial and temporal dimensions more effectively and complete the SST prediction by spatiotemporal fusion. The experiments prove that the proposed model greatly improves the prediction accuracy and is an effective model for SST prediction.

The accurate temperature background field plays a vital role in the numerical prediction of sea surface temperature (SST). At present, the SST background field is mainly derived from multi-source data fusion, including satellite SST data and in situ data from marine stations, buoys, and voluntary observing ships. The characteristics of satellite SST data are wide coverage but low accuracy, whereas the in situ data have high accuracy but sparse distribution. For obtaining a more accurate temperature background field and realizing the fusion of measured data with satellite data as much as possible, we propose a sparse data-extended fusion method to predict SST in this paper. By using this method, the actual observed sites and buoys data in the East China Sea area are fused with Advanced Very High Resolution Radiometer (AVHRR) Pathfinder Version 5.0 SST data. Furthermore, the temperature field in the study area were predicted by using Long Short-Term Memory (LSTM) and Gate Recurrent Unit (GRU) deep learning methods, respectively. Finally, we obtained the results by traditional prediction methods to verify them. The experimental results show that the method we proposed in this paper can obtain more accurate prediction results, and effectively compensate for the uncertainty caused by the parameterization of ocean dynamic process, the discrete method, and the error of initial conditions.