Accurate daily streamflow forecasting is crucial for effective flood control and water management. Decomposition ensemble models have proven to be effective in daily streamflow forecasting. However, it is important to...
详细信息
Accurate daily streamflow forecasting is crucial for effective flood control and water management. Decomposition ensemble models have proven to be effective in daily streamflow forecasting. However, it is important to recognize that the performance of streamflow forecasting within the decomposition ensemble models is significantly influenced by the choice of forecast model used. Therefore, a thoughtful selection of streamflow forecast models can greatly enhance the performance of decomposition ensemble models. Nevertheless, simple forecast models, such as long short-term memory, are susceptible to gradient disappearance or explosion, resulting in suboptimal peak streamflow forecasting performance, especially when dealing with long streamflow series. This limitation ultimately hinders the predictive accuracy of the decomposition ensemble model. Furthermore, while single-model forecasting (SF) schemes have demonstrated effectiveness, there has been limited exploration of the potential for coupling VMD and encoder-decoder framework in mid- and long-term daily streamflow forecasting. To address these issues and further enhance the capability of common VMD-LSTM model in mid and long-term streamflow forecasting, we have developed a novel and efficient forecast model VMD-LSTM-ED by integrating VMD and LSTM with a robust encoder-decoder structure in the SF scheme. Our proposed method was tested in the Fish River Basin in Maine, USA. To demonstrate its superiority, we also compared VMD-LSTM-ED against LSTM-ED, VMD-LSTM, and LSTM models. The results indicate that the proposed VMD-LSTM-ED exhibits outstanding performance across all aspects, with its advantage becoming more prominent as lead times increase, particularly in peak streamflow forecasting. For example, compared to VMD-LSTM, the Nash-Sutcliffe Efficiency (NSE) of VMD-LSTM-ED increased by 2.91%, 8.45%, 20.2% and 51.2% respectively, at lead times of 1, 3, 5, and 7 days;thus significantly enhancing the long-term predict
Accurate diagnosis of the state-of-health (SOH) of the lithium-ion battery is crucial for its safe and reliable operation. In this article, a two-level battery health diagnosis model is proposed using relaxation volta...
详细信息
Accurate diagnosis of the state-of-health (SOH) of the lithium-ion battery is crucial for its safe and reliable operation. In this article, a two-level battery health diagnosis model is proposed using relaxation voltage. First, the health features of the relaxation voltage sequence are extracted using the autoencoder based on the encoder-decoder framework, and the Gaussian mixture model (GMM) is used to cluster distinct aging levels of the battery, thereby enabling a preliminary diagnosis of battery SOH. Then, a novel Gaussian mixture ensemble learning (GMEL) method is presented that leverages prior knowledge for accurate diagnosis of battery SOH. Furthermore, the performance of the ensemble model is enhanced by the sequential model-based algorithm configuration (SMAC) algorithm to optimize the hyperparameters, resulting in mean-absolute-error (MAE) and root-mean-square-error (RMSE) of 0.736% and 1.013%, respectively. In addition, a data reconstruction model is developed using the encoder-decoder framework to address the challenge of obtaining complete relaxation voltage sequences in the real world. Utilizing only 4 min of incomplete relaxation voltage data, the presented model achieves the MAE of 1.265% and RMSE of 1.681% in diagnosing the battery SOH. Finally, the superiority of the proposed method is verified by several sets of experiments.
Numerous IoT applications have emerged in human healthcare area with advances in wearable electronics. Multiple physical and physiological data bearing strong spatial-temporal characteristic collected by the wearable ...
详细信息
Numerous IoT applications have emerged in human healthcare area with advances in wearable electronics. Multiple physical and physiological data bearing strong spatial-temporal characteristic collected by the wearable sensors is sent to the smartphones where it is aggregated and transferred to back-end applications for further processing. Analyzing the multivariate time series is of great importance yet very challenging as it is affected by many complex factors, i.e., dynamic spatial-temporal correlations and external factors. In this paper, we propose an attention-based encoder-decoder framework for multi-sensory time-series analytic. It consists of four parts: data collection, data mining, time-series analytic part and user interaction. A temporal-attention based encoder-decoder model is proposed to make a long-term prediction of multiple time series to realize the real-time user interaction. The proposed model uses the LSTM model to learn the long-term dependence of the time series related to certain motion sequence. The attention mechanism connects the encoder and the decoder to make long-term predictions for future time series. Through extensive experiments, the proposed model has achieved better results in short-term and long-term predictions compared with the state of art methods. An activity recognition algorithm based on LSTM is also proposed in this framework to identify daily human activities and sports activities accurately. Through five-fold and ten-fold cross-validation strategies and comparison with six baseline machine learning models, the activity recognition algorithm has a recognition rate of 98.89% and 99.28% for human activity.
Anomaly detection in smart grid is critical to enhance the reliability of power systems. Excessive manpower has to be involved in analyzing the measurement data collected from intelligent motoring devices while perfor...
详细信息
Anomaly detection in smart grid is critical to enhance the reliability of power systems. Excessive manpower has to be involved in analyzing the measurement data collected from intelligent motoring devices while performance of anomaly detection is still not satisfactory. This is mainly because the inherent spatio-temporality and multi-dimensionality of the measurement data cannot be easily captured. In this paper, we propose an anomaly detection model based on encoder-decoder framework with recurrent neural network (RNN). In the model, an input time series is reconstructed and an anomaly can be detected by an unexpected high reconstruction error. Both Manhattan distance and the edit distance are used to evaluate the difference between an input time series and its reconstructed one. Finally, we validate the proposed model by using power demand data from University of California, Riverside (UCR) time series classification archive and IEEE 39 bus system simulation data. Results from the analysis demonstrate that the proposed encoder-decoder framework is able to successfully capture anomalies with a precision higher than 95%.
Aiming at the situation that complex natural scene text is difficult to recognize a scene text recognition method based on an encoder-decoder framework is proposed. The method converts the natural text recognition int...
详细信息
Aiming at the situation that complex natural scene text is difficult to recognize a scene text recognition method based on an encoder-decoder framework is proposed. The method converts the natural text recognition into a sequence mark by combining the connection time classification (CTC) and attention mechanism under the encoder-decoder framework, in order to overcome the problem of character segmentation, using the correlation between image and text sequence. First of all, a convolutional neural network (CNN) is used to generate an ordered feature sequence from the entire word image. Then, the generated feature sequence is feature-coded using the bidirectional long short-term memory (Bi-LSTM) network. Finally, an integrated module of the CTC and attention mechanism is designed to decode and output the text sequence. The experiments show that compared with the comparison method, the recognition accuracy of the method is improved obviously.
Although automatic license plate recognition (ALPR) has been studied for decades, the final recognition result can be accurate only if the license plate is detected and the standard format is unambiguous. However, sin...
详细信息
Although automatic license plate recognition (ALPR) has been studied for decades, the final recognition result can be accurate only if the license plate is detected and the standard format is unambiguous. However, since an image may contain license plates with different formats and scales, license plate detection and standard format classification may fail. In this study, a new ALPR codec framework named EDF-LPR is presented. As for the encoder, at the first stage, candidate license plate characters are detected and recognised directly without considering the format of license plate, and candidate regions of characters are extracted by density-based spatial clustering of applications with noise-like algorithm;at the second stage, poor regions are processed by tilt correction and scale normalisation to obtain more accurate candidate characters. As for the decoder, a sequence learning model is trained to convert each unordered coded sequence into a sequence composed of marks that indicate a way to construct the final result string. Experiments are designed to evaluate the performance of EDF-LPR on both detection rate and recognition rate. The experimental results on public datasets show that the detection rate and recognition rate are 99.51 and 95.3%, respectively, at about 40 fps.
Jaywalker-vehicle (J-V) conflicts at mid-blocks without crossing facilities in China are frequent and hazardous. Due to the unexpected and sudden nature of jaywalking activity, it is crucial to develop predictive mode...
详细信息
Jaywalker-vehicle (J-V) conflicts at mid-blocks without crossing facilities in China are frequent and hazardous. Due to the unexpected and sudden nature of jaywalking activity, it is crucial to develop predictive models for J-V conflicts to offer pre-conflict warnings for road users. This study introduces a novel encoder-decoder framework that utilizes multi-source data to predict J-V conflict severity. We define three encoders to represent three types of input data, (1) J-V interaction encoder (Bi-LSTM), (2) jaywalker motion encoder (Bi-LSTM) and (3) back-ground information encoder (MLP). Subsequently, features extracted by these three encoders are concatenated and transferred to the conflict severity decoder (MLP) to obtain the predicted severity *** further conduct a case study using the surveyed video data at three mid-blocks without crossing facilities in Nanjing, China. The experimental results indicate that, compared to classical models, our proposed encoder-decoder (Proposed ED) model exhibits the best and stable predictive metrics. Furthermore, the results of the ablation study suggest that the incorporation of background information significantly enhances the four evalu-ative metrics of the Proposed ED model, with an average improvement of 24.291%. Additionally, the results of transferability analysis suggest that, when the ratio of added samples from the new mid-block reaches 40% to 50%, the predictive metrics of the updated models could stabilize at around 80% to 95%, indicating a notably good performance. Eventually, we derive several practical suggestions from the above findings, in order to help with J-V conflict prediction and jaywalking safety improvement.
Human motion prediction is an important and challenging task in computer vision with various applications. Recurrent neural networks (RNNs) and convolutional neural networks (CNNs) have been proposed to address this c...
详细信息
Human motion prediction is an important and challenging task in computer vision with various applications. Recurrent neural networks (RNNs) and convolutional neural networks (CNNs) have been proposed to address this challenging task. However, RNNs exhibit their limitations on long-term temporal modelling and spatial modelling of motion signals. CNNs show their inflexible spatial and temporal modelling capability that mainly depends on a large convolutional kernel and the stride of convolutional operation. Moreover, those methods predict multiple future poses recursively, which easily suffer from noise accumulation. The authors present a new encoder-decoder framework based on the residual convolutional block with a small filter to predict future human poses, which can flexibly capture the hierarchical spatial and temporal representation of the human motion signals from the motion capture sensor. Specifically, the encoder is stacked by multiple residual convolutional blocks to hierarchically encode the spatio-temporal features of previous poses. The decoder is built with two fully connected layers to automatically reconstruct the spatial and temporal information of future poses in a non-recursive manner, which can avoid noise accumulation that differs from prior works. Experimental results show that the proposed method outperforms baselines on the Human3.6M dataset, which shows the effectiveness of the proposed method. The code is available at https://***/lily2lab/residual_prediction_network.
The ongoing COVID-19 pandemic has dramatically changed people's daily lives. A robust forecasting model for COVID-19 infections is essential for governments and institutions to plan timely and perform accurate int...
详细信息
ISBN:
(纸本)9781450384469
The ongoing COVID-19 pandemic has dramatically changed people's daily lives. A robust forecasting model for COVID-19 infections is essential for governments and institutions to plan timely and perform accurate interventions. Mainstream solutions for COVID-19 prediction fit reported data only by considering observed cases. However, the neglected facts that positive samples are incomplete and many facts of the novel disease are unknown may be prone to cause severe error accumulation, especially in long-term predictions. To fully understand the spreading patterns of the virus, we propose an encoder-decoder framework: (i) in the encoder we embed historical case data into multiple expose-infection ranges and learn message passing between time slices and across ranges with coarse-grained human mobility data incorporated;(ii) in the decoder, we decode the embedded features based on reported cases as well as deaths to jointly consider the effect of both observed and hidden data. We model the spreading of disease in over 60 counties of California and New York, which are two of the most metropolitan areas in the US. The proposed framework significantly outperforms state-of-the-art baselines on JHU COVID-19 dataset on both weekly prediction and daily prediction tasks. We design detailed ablation studies to verify the effectiveness of each key module and find the model not only works with the assistance of mobility data but also with purely cases and deaths, which implies its broad application scenarios.
Many consumers resort to others' assessments in product reviews for decision making, while their time is limited to deal with many reviews. Therefore, an expert's review which contains all important features i...
详细信息
ISBN:
(纸本)9781538693803
Many consumers resort to others' assessments in product reviews for decision making, while their time is limited to deal with many reviews. Therefore, an expert's review which contains all important features in user-generated reviews, is strongly expected. In this paper, we study "how to generate expert's review from a large number of user-generated reviews (i.e., the crowds)." It can be implemented by text summarization, which mainly has two types of the extractive and the abstractive approaches. However, the former may generate redundant and incoherent summaries, while the latter cannot deal with long sequences. Moreover, both approaches usually neglect the sentiment information. To address the above issues, we propose a novel Expert Review Generation model to integrate a multi-attention mechanism with the encoder-decoder framework. We design a comprehensive preprocess strategy to identify the important sentences while keeping users' sentiment in original reviews, and use them as the input of the encoder-decoder generation model, so as to generate non-redundant and coherent summaries. Experimental results in two real world data sets (Idebate and Rotten Tomatoes) demonstrate that our model performs well in expert review generation.
暂无评论