Accurate prediction of water quality contributes to the intelligent management of water resources. Water quality indices have time series characteristics and nonlinearity, but the existing models only focus on the forward time series when long short-term memory (LSTM) is introduced and do not consider the parallel computation on the model. Owing to this, a new neural network called LSTM-multihead attention (LMA) was constructed to predict water quality, using long short-term memory to process time series data and multihead attention for parallel computing and extracting feature information. Additionally, water quality indices have the issues of multiple data types and complex data correlations, as well as missing data and abnormal data problems in water quality data. In order to solve these problems, this study proposes a water quality prediction model called GRA-LMA-based linear interpolation, gray relational analysis and LMA. Two experiments are carried out to verify the predictive performance of the GRA-LMA with the water quality data of the Huaihe River Basin as a case study sample. The first experiment focuses on data processing, including the processing of missing data and abnormal data of water quality data, and the correlation analysis of water quality indices. Linear interpolation is adapted to process the missing data, while a combination of boxplot and histogram is adopted to analyze and eliminate the abnormal data, which is then repaired the abnormal data with linear interpolation. The gray relational analysis is adopted to calculate the correlation between different water quality indices, and water quality indices with high correlation are retained to determine the input variables of the water quality prediction model. The data processing results demonstrate that repairs can be made using linear interpolation without altering the pattern of data change and the model by using the gray relational analysis to reduce the quantity of data it needs as input. In the second experiment, the predictive capacity of GRA-LMA and existing models such as backpropagation neural network (BP), recurrent neural network (RNN), long short-term memory (LSTM), and gate recurrent unit (GRU) was evaluated and compared using different numerical and graphical performance evaluation metrics. Comparative experimental results show that the mean square error of pH, dissolved oxygen, chemical oxygen demand, ammonia nitrogen, electrical conductivity, turbidity, total phosphorus, and total nitrogen of GRA-LMA is reduced to 0.05890, 0.40196, 0.32454, 0.04368, 14.71003, 8.13252, 0.01558, and 0.14345. The results indicate that GRA-LMA has superior adaptability for predicting various water quality indices and can significantly lower the induced prediction error.

OBJECTIVE: Deep learning (DL) is a state-of-the-art technique for developing artificial intelligence in various domains and it improves the performance of natural language processing (NLP). Therefore, we aimed to develop a DL-based NLP model that classifies the status of bone metastasis (BM) in radiology reports to detect patients with BM.
METHODS: The DL-based NLP model was developed by training long short-term memory using 1,749 free-text radiology reports written in Japanese. We adopted five-fold cross-validation and used 200 reports for testing the five models. The accuracy, sensitivity, specificity, precision, and area under the receiver operating characteristics curve (AUROC) were used for the model evaluation.
RESULTS: The developed model demonstrated classification performance with mean ± standard deviation of 0.912 ± 0.012, 0.924 ± 0.029, 0.901 ± 0.014, 0.898 ± 0.012, and 0.968 ± 0.004 for accuracy, sensitivity, specificity, precision, and AUROC, respectively.
CONCLUSIONS: The proposed DL-based NLP model may help in the early and efficient detection of patients with BM.

The use of models capable of forecasting the production of photovoltaic (PV) energy is essential to guarantee the best possible integration of this energy source into traditional distribution grids. Long Short-Term Memory networks (LSTMs) are commonly used for this purpose, but their use may not be the better option due to their great computational complexity and slower inference and training time. Thus, in this work, we seek to evaluate the use of neural networks MLPs (Multilayer Perceptron), Recurrent Neural Networks (RNNs), and LSTMs, for the forecast of 5 min of photovoltaic energy production. Each iteration of the predictions uses the last 120 min of data collected from the PV system (power, irradiation, and PV cell temperature), measured from 2019 to mid-2022 in Maceió (Brazil). In addition, Bayesian hyperparameters optimization was used to obtain the best of each model and compare them on an equal footing. Results showed that the MLP performs satisfactorily, requiring much less time to train and forecast, indicating that they can be a better option when dealing with a very short-term forecast in specific contexts, for example, in systems with little computational resources.

Carbon trading price (CTP) prediction accuracy is critical for both market participants and policymakers. As things stand, most previous studies have only focused on one or a few carbon trading markets, implying that the models\' universality is insufficient to be validated. By employing a case study of all carbon trading markets in China, this study proposes a hybrid point and interval CTP forecasting model. First, the Pearson correlation method is used to identify the key influencing factors of CTP. The original CTP data is then decomposed into multiple series using complete ensemble empirical mode decomposition with adaptive noise. Following that, the sample entropy method is used to reconstruct the series to reduce computational time and avoid overdecomposition. Following that, a long short-term memory method optimized by the Adam algorithm is established to achieve the point forecasting of CTP. Finally, the kernel density estimation method is used to predict CTP intervals. On the one hand, the results demonstrate the proposed model\'s validity and superiority. The interval prediction model, on the other hand, reflects the uncertainty of market participants\' behavior, which is more practical in the operation of carbon trading markets.

Real-time prediction of the state of the river itself and the degree of its benefit to the people is the leading way to achieve human-water harmony. Using the indicator scoring method as the evaluation method, we used the river evaluation data and results with time series characteristics as features and labels and applied the concept of transfer learning to Long Short-Term Memory to establish six subsystems, including water safety, water quality, economic contribution, water ecology, water management and water culture, to conduct a real-time rolling evaluation simulation study on the degree of river happiness in the Jiangsu section of the Huaihe River Basin in China. The empirical results show that the maximum Root Mean Square Error (RMSE) of the training set and test set of each system is 0.0226, and the lowest coefficient of determination R2 is 0.9011, which proves that the model fits well, according to which the relevant data of the watershed in June 2022 are brought in, and the evaluation result is obtained as 89.77 points. The overall trend is good, but a certain tendency to fall back at the level of economic contribution can be found, and the reasons are analyzed objectively.

Current designs of advanced driving assistance systems (ADAS) mainly developed uniform collision warning algorithms, which ignore the heterogeneity of driving behaviors, thus lead to low drivers\' trust in. To address this issue, developing personalized driving assistance algorithms is a promising approach. However, current personalization systems were mainly implemented through manually adjusting warning trigger thresholds, which would be less feasible for overall drivers as certain domain expertise is required to set personal thresholds accurately. Other personalization techniques exploited individual drivers\' data to build personalized models. Such approach could learn personal behavior but requires impractical large-scale individual data collections. To fill up the gaps, self-adaptive algorithms for personalized forward collision warning (FCW) based on federated learning were proposed in this study. A baseline model was developed by long short-term memory (LSTM) for FCW. Federated learning framework was then introduced to collect knowledge from multiple drivers with privacy preserving. Specifically, a general cloud server model was trained by collecting updated parameters from individual vehicle server models rather than collecting raw data. Besides, a driver-specific batch normalization (BN) layer was added into each vehicle server model to address the heterogeneity of driving behaviors. Experiments show empirically that the proposed federated-based personalized models with the BN layer showed to have the best performance. The average modeling accuracy has reached 84.88% and the performance is comparable to conventional total data collection training approach, where the additional BN layer could increase the accuracy by 3.48%. Finally, applications of the proposed framework and its further investigations have been discussed.

This study was conducted using a drone with advanced mobility to develop a unified sensor and communication system as a new platform for in situ atmospheric measurements. As a major cause of air pollution, particulate matter (PM) has been attracting attention globally. We developed a small, lightweight, simple, and cost-effective multi-sensor system for multiple measurements of atmospheric phenomena and related environmental information. For in situ local area measurements, we used a long-range wireless communication module with real-time monitoring and visualizing software applications. Moreover, we developed four prototype brackets with optimal assignment of sensors, devices, and a camera for mounting on a drone as a unified system platform. Results of calibration experiments, when compared to data from two upper-grade PM2.5 sensors, demonstrated that our sensor system followed the overall tendencies and changes. We obtained original datasets after conducting flight measurement experiments at three sites with differing surrounding environments. The experimentally obtained prediction results matched regional PM2.5 trends obtained using long short-term memory (LSTM) networks trained using the respective datasets.

Basal/acetazolamide brain perfusion single-photon emission computed tomography (SPECT) has been used to evaluate functional hemodynamics in patients with carotid artery stenosis. We aimed to develop a deep learning model as a support system for interpreting brain perfusion SPECT leveraging unstructured text reports.
In total, 7345 basal/acetazolamide brain perfusion SPECT images and their text reports were retrospectively collected. A long short-term memory (LSTM) network was trained using 500 randomly selected text reports to predict manually labeled structured information, including abnormalities of basal perfusion and vascular reserve for each vascular territory. Using this trained LSTM model, we extracted structured information from the remaining 6845 text reports to develop a deep learning model for interpreting SPECT images. The model was based on a 3D convolutional neural network (CNN), and the performance was tested on the other 500 cases by measuring the area under the receiver-operating characteristic curve (AUC). We then applied the model to patients who underwent revascularization (n = 33) to compare the estimated output of the CNN model for pre- and post-revascularization SPECT and clinical outcomes.
The AUC of the LSTM model for extracting structured labels was 1.00 for basal perfusion and 0.99 for vascular reserve for all 9 brain regions. The AUC of the CNN model designed to identify abnormal perfusion was 0.83 for basal perfusion and 0.89 for vascular reserve. The output of the CNN model was significantly improved according to the revascularization in the target vascular territory, and its changes in brain territories were concordant with clinical outcomes.
We developed a deep learning model to support the interpretation of brain perfusion SPECT by converting unstructured text reports into structured labels. This model can be used as a support system not only to identify perfusion abnormalities but also to provide quantitative scores of abnormalities, particularly for patients who require revascularization.