Single LSTM Research Articles

Energy price prediction based on decomposed price dynamics: A parallel neural network approach

The growing uncertainties of the world due to geographic tensions, and weather conditions challenge the traditional method to achieve a reliable price prediction of energy future for both asset pricing and risk management. Following the initial success of deep learning models in energy price prediction, we attempt to establish a better architecture of neural networks to improve the prediction accuracy. We propose a novel Parallel Hybrid Neural Network (PHNN) model that utilizes independent sub-networks to effectively capture the distinct features of various sequences. Empirical results demonstrate that the PHNN model exhibits a significant performance enhancement of 16.68%, 14.09%, and 2.34% over the EMD–LSTM, the informer model, and the single LSTM model, respectively. In particular, the PHNN outperforms the single LSTM, which is trained on the same inputs, by 2.34% overall while by a remarkable 4.11% during event periods. This suggests that the PHNN derives notable benefits from its distinct architecture, particularly during the initial phase of extreme events characterized by significant price trend changes. Additionally, the study explores the potential benefits of incorporating additional event-tracking indicators for energy futures price forecasting. However, the findings suggest that these indicators may not consistently provide effective complementary information.

Application of Singular Spectrum Analysis (SSA) Decomposition in Artificial Neural Network (ANN) Forecasting

Over time, various forecasting methods have been introduced. An example is the Hybrid model. This model can enhance the forecast accuracy compared to a single model. The Hybrid Singular Spectrum Analysis (SSA)-Artificial Neural Network (ANN) model combines the concepts of decomposition and forecasting. The Hybrid SSA-ANN forecasting works through two stages. Firstly, SSA decomposes the data into trend, seasonal, noise, and residue components. Secondly, the decomposed data is predicted using the ANN model, specifically the LSTM and GRU models. The Hybrid SSA-ANN model has been proven to improve forecasting accuracy. The Hybrid SSA-LSTM model improves the forecast accuracy by 78% compared to the single LSTM forecasting model. This can be seen from the respective RMSE values of 4.36 changing to 0.97 and MAPE values of 5.2% changing to 1.16%. Similarly, the Hybrid SSA-GRU model improves the forecast accuracy by 79% compared to the single GRU forecasting model. This can be observed from the respective RMSE values of 4.86 changing to 1.01 and MAPE values of 6.33% changing to 1.36%. In a case study using weekly data of crude oil's opening prices, the application of SSA decomposition can enhance the forecast accuracy by 78-79% in ANN forecasting

Runoff Prediction in Different Forecast Periods via a Hybrid Machine Learning Model for Ganjiang River Basin, China

Accurate forecasting of monthly runoff is essential for efficient management, allocation, and utilization of water resources. To improve the prediction accuracy of monthly runoff, the long and short memory neural networks (LSTM) coupled with variational mode decomposition (VMD) and principal component analysis (PCA), namely VMD-PCA-LSTM, was developed and applied at the Waizhou station in the Ganjiang River Basin. The process begins with identifying the main forecasting factors from 130 atmospheric circulation indexes using the PCA method and extracting the stationary components from the original monthly runoff series using the VMD method. Then, the correlation coefficient method is used to determine the lag of the above factors. Lastly, the monthly runoff is simulated by combining the stationary components and key forecasting factors via the LSTM model. Results show that the VMD-PCA-LSTM model effectively addresses the issue of low prediction accuracy at high flows caused by a limited number of samples. Compared to the single LSTM and VMD-LSTM models, this comprehensive approach significantly enhances the model’s predictive accuracy, particularly during the flood season.

Adaptive prediction for effluent quality of wastewater treatment plant: Improvement with a dual-stage attention-based LSTM network

The accurate effluent prediction plays a crucial role in providing early warning for abnormal effluent and achieving the adjustment of feedforward control parameters during wastewater treatment. This study applied a dual-staged attention mechanism based on long short-term memory network (DA-LSTM) to improve the accuracy of effluent quality prediction. The results showed that input attention (IA) and temporal attention (TA) significantly enhanced the prediction performance of LSTM. Specially, IA could adaptively adjust feature weights to enhance the robustness against input noise, with R2 increased by 13.18%. To promote its long-term memory ability, TA was used to increase the memory span from 96 h to 168 h. Compared to a single LSTM model, the DA-LSTM model showed an improvement in prediction accuracy by 5.10%, 2.11%, 14.47% for COD, TP, and TN. Additionally, DA-LSTM demonstrated excellent generalization performance in new scenarios, with the R2 values for COD, TP, and TN increasing by 22.67%, 20.06%, and 17.14% respectively, while the MAPE values decreased by 56.46%, 63.08%, and 42.79%. In conclusion, the DA-LSTM model demonstrated excellent prediction performance and generalization ability due to its advantages of feature-adaptive weighting and long-term memory focusing. This has forward-looking significance for achieving efficient early warning of abnormal operating conditions and timely management of control parameters.

AIRS: A QGIS plugin for time series forecasting using deep learning models

Time series forecasting, particularly when applied to geospatial data, serves as an essential tool for accurate observation and prediction of environmental and spatial occurrences. Recently, the integration of deep learning models into forecasting processes has garnered increased importance. Deep learning methods offer enhanced capabilities for identifying complex patterns within geospatial datasets, leading to more accurate forecasts. Nevertheless, there is a critical need for analyzing and identifying effective deep learning models in order to assure the accuracy of forecasting outcomes. This study presents an open-source QGIS plugin named AIRS (Artificial Intelligence forecasting Remote Sensing). This plugin allows time series forecasting using five deep learning models (i.e., FFNN, single LSTM, stacked LSTM, BiLSTM, and Conv-LSTM) and provides a user-friendly tool permitting data processing, model building and training, future prediction, accuracy analysis, and results visualization and saving. AIRS is written in Python using QGIS internal and external packages, with an easy-to-use GUI interface.

DLSTM-MSF: Distributed LSTM Models for Multimedia Streaming Workload Forecasting Based on Kafka Environment

This paper introduces DLSTM-MSF, a distributed approach designed to address the challenge of demand forecasting in multimedia streaming workloads. DLSTM-MSF leverages the power of multi-LSTM networks, each tailored to predict data demand for a specific type of multimedia streaming workload. The central problem addressed in this research is the accurate prediction of workload demand in a dynamic and diverse multimedia streaming environment. To achieve specialization, the training time series set for each LSTM network comprises examples with targets belonging exclusively to the workload type it is designed to predict. This specialization ensures that each LSTM network becomes proficient at capturing the unique demand patterns associated with its designated workload category. The methodology of the proposed approach is based on building the best forecasting model for each multimedia streaming workload type by exploring various combinations of LSTM hyper-parameters using the grid search method. This enables the proposed approach to effectively capture nonlinear patterns in time series data. Furthermore, the implementation of DLSTM-MSF incorporates Apache Kafka for online demand prediction, utilizing the best-developed model for each workload type. Experimental evaluations of DLSTM-MSF compare the performance of two ensemble-learning LSTM models (Ensemble V1 and Ensemble V2) with a single LSTM model. The results unequivocally highlight the superiority of Ensemble V1, with reductions of 71.85% and 74.88% in RMSE and MAE values, respectively, compared to the single LSTM model. Index Terms— Multimedia streaming, LSTM, Ensemble learning, Forecasting, Workload Demand, Big data.

Isolated Forest-Based Prediction of Container Resource Load Extremes

Given the wide application of container technology, the accurate prediction of container CPU usage has become a core aspect of optimizing resource allocation and improving system performance. The high volatility of container CPU utilization, especially the uncertainty of extreme values of CPU utilization, is challenging to accurately predict, which affects the accuracy of the overall prediction model. To address this problem, a container CPU utilization prediction model, called ExtremoNet, which integrates the isolated forest algorithm, and classification sub-models are proposed. To ensure that the prediction model adequately takes into account critical information on the CPU utilization’s extreme values, the isolated forest algorithm is introduced to compute these anomalous extreme values and integrate them as features into the training data. In order to improve the recognition accuracy of normal and extreme CPU utilization values, a classification sub-model is used. The experimental results show that, on the AliCloud dataset, the model has an R2 of 96.51% and an MSE of 7.79. Compared with the single prediction models TCN, LSTM, and GRU, as well as the existing combination models CNN-BiGRU-Attention and CNN-LSTM, the model achieves average reductions in the MSE and MAE of about 38.26% and 23.12%, proving the effectiveness of the model at predicting container CPU utilization, and provides a more accurate basis for resource allocation decisions.

Track Circuits Fault Diagnosis Method Based on the UNet-LSTM Network (ULN)

As a commonly used mode of transportation in people’s daily lives, the normal operation of railway transportation is crucial. The track circuit, as a key component of the railway transportation system, is prone to malfunctions due to environmental factors. However, the current method of inspecting track circuit faults still relies on the experience of on-site personnel. In order to improve the efficiency and accuracy of fault diagnosis, we propose to establish an intelligent fault diagnosis system. Considering that the fault data are a one-dimensional time series, this paper presents a fault diagnosis method based on the UNet-LSTM network (ULN). The LSTM network is established on the basis of fault data and used for ZPW-2000A track circuit fault diagnosis. However, the use of a single LSTM network has a high error rate in the common fault diagnosis of track circuits. Therefore, this paper proposes a feature extraction method based on the UNet network. This method is used to extract the features of the original data and then input them into the LSTM network for fault diagnosis. Through experiments with on-site fault data, it has been verified that this method can accurately classify seven common track circuit faults. Finally, the superiority of the method is verified by comparing it with other commonly used fault classification methods.

A CNN-LSTM based deep learning model with high accuracy and robustness for carbon price forecasting: A case of Shenzhen's carbon market in China

Accurately predicting carbon trading prices using deep learning models can help enterprises understand the operational mechanisms and regulations of the carbon market. This is crucial for expanding the industries covered by the carbon market and ensuring its stable and healthy development. To ensure the accuracy and reliability of the predictions in practical applications, it is important to evaluate the model's robustness. In this paper, we built models with different parameters to predict carbon trading prices, and proposed models with high accuracy and robustness. The accuracy of the models was assessed using traditional survey indicators. The robustness of the CNN-LSTM model was compared to that of the LSTM model using Z-scores. The CNN-LSTM model with the best prediction performance was compared to a single LSTM model, resulting in a 9% reduction in MSE and a 0.0133 shortening of the Z-score range. Furthermore, the CNN-LSTM model achieved a level of accuracy comparable to other popular models such as CEEMDAN, Boosting, and GRU. It also demonstrated a training speed improvement of at least 40% compared to the aforementioned methods. These results suggest that the CNN-LSTM enhances model resilience. Moreover, the practicality of using Z-score to evaluate model robustness is confirmed.

A Stock Index Prediction Model Based on LSTM: Taking New Energy Vehicles as an Example

Stock indices can partially reflect the overall price level and variations of various stocks in the market. Therefore, analyzing the stock index of new energy vehicles aids in predicting future market trends. However, stock prices, as time series data, display traits of nonlinearity and high noise. This study proposes employing the denoising method of Singular Spectrum Analysis (SSA) on the CNI New Energy Vehicles Index (399417) and constructing an SSA-LSTM model. Comparative results with a single LSTM model demonstrate that the evaluation indicators R2 of the SSA-LSTM model increase by 0.06, RMSE decreases by 54.37, and MAE decreases by 49.49. These findings suggest that SSA can enhance the predictive accuracy of the model.

Incorporating multiple grid-based data in CNN-LSTM hybrid model for daily runoff prediction in the source region of the Yellow River Basin

Study regionThe Source Region of the Yellow River Basin (SRYRB), China Study focusTo improve daily runoff prediction accuracy in data-scarce areas, this study focuses on incorporating multiple grid-based data (precipitation, EVI, soil moisture (SM)) to drive the CNN-LSTM hybrid model. The spatial features of precipitation and underlying surface of the basin can be extracted by CNN, while the temporal features of the input data series can be captured by the LSTM. The hybrid model is compared with the single models (CNN, LSTM), and hybrid model performances under different driven data are also investigated. New hydrological insights for the regionDriven by the in-situ precipitation, grid-based precipitation (GPM) and SM data, the CNN-LSTM hybrid model achieved the best prediction result with NSE of 0.834, outperforming the single LSTM model (NSE=0.510) and the CNN model (NSE=0.612). It indicates that the hybrid model captures the spatiotemporal change features of precipitation and underlying surface of the basin. When using only GPM and SM data as input, the hybrid model achieved comparable result with NSE of 0.827. It implies that GPM could serve as a good alternative of in-situ precipitation and SM could provide additional value to improve prediction. This study highlights the value of using multiple grid-based data to drive the hybrid model, which provides new insights into runoff prediction in data-scarce regions.

Exploring the Impact of Digital Shared Services in Finance on Corporate Treasurers’ Job Anxiety

Abstract It has become a trend for enterprises to establish financial shared service centers for digital financial transformation, but the emotional impact of this transformation on the original financial staff of enterprises has been neglected. In this paper, we collect relevant intranet texts of enterprises, mine the emotions of the collected texts through the LDA theme algorithm, use the improved LSTM model to classify and identify the emotions of the collected texts, and construct the LDA-BiLSTM sentiment analysis model to analyze the anxiety condition of the financial staff of enterprises. The efficiency of LDA can be improved by using the weighted median to deal with outliers. Two LSTM networks are combined to form a BiLSTM network to improve the problem of inaccurate judgment of sentiment indicators by a single LSTM. Finally, the model is used to empirically analyze the emotions of financial staff before and after the transformation of an enterprise’s financial digital service. Negative feelings that are linked to anxiety have an accuracy rate of 95.23%. The frequency of separation after the transition was as high as 0.5, and the frequency of dismissal was as high as 0.423. Overall sentiment scores were lower than 0.7. After the transition, the number of people worried about negative feelings related to anxiety rose from 32% to 64%. Finance professionals are experiencing a significant increase in anxiety due to the post-transition of the enterprise.

Integrated prediction of green bond return under the dual risks of climate change and energy crisis

Prediction of bond return is a classic problem in financial area, providing an important basis for portfolio construction and risk management. The sustainable investment attribute of green bonds has been favored by investors, so that green bonds have become an important component for major asset allocation. However, due to the specific investment focus of green bonds, investors’ return expectations are influenced not only by traditional corporate bond factors, but also by related factors such as climate change and energy transition. Against the backdrop of increasingly severe climate risks and the global energy crisis, this paper analyses the volatility characteristics of China’s green bonds at multiple time scales, and introduces exogenous variables such as returns of the alternative financial assets, climate risks and returns of energy markets for prediction. Based on the LSTM model, the volatility of green bond yield at different time scales is separately predicted using optimal exogenous variable before integration. It is found that the new integrated prediction model can significantly improve the forecasting performance compared to traditional single LSTM models and simple decomposition-integrated models. Further, both climate risks and energy markets variables have a significant improvement effect on predicting green bond in low-frequency item, while energy markets variables also have a better predictive effect on trend items. Building on the use of only LSTM model, it could be further enhanced by integrating more algorithms to select the best single model for each component, further improve the prediction accuracy and provide a more effective quantitative tool for investment decision-making and risk management in related fields.

Enhancing Stock Price Prediction Method Based on CNN-LSTM hybrid model

Stock price prediction means using the previous data to predict future trend. With the development of deep learning technologies in artificial intelligence area, many works try to use single CNN or single LSTM to make prediction. However, both these two models have their own weakness and strengths. In this paper, we try to make a CNN-LSTM hybrid model to achieve a better performance on stock price prediction. It proposes that the CNN-LSTM hybrid model predicts stock prices using RMSE to measure higher accuracy than a single CNN and a single LSTM model. The experiment shows a great ability on generalization and provide reliable results, which can apply to other stocks to make prediction and give suggestions on investing strategy. The CNN-LSTM model can play the respective advantages of CNN and LSTM model to improve the accuracy of stock price prediction and has better performance in both short-term prediction and long-term trend.

Simultaneous temporal and spectral evolution of ultrafast optical pulse propagation using a single bidirectional LSTM network

Numerical simulations of optical pulse evolution are often a cumbersome task, mostly due to the presence of nonlinearities, and standard numerical approaches like splitting methods can be too complex and resource-hungry to compute them. However, under certain conditions, deep learning (DL) techniques can provide accurate alternative approaches to predict such propagation behavior. Here, we show a bidirectional long short-term memory (BiLSTM) network that predicts both temporal and spectral evolution of a short optical pulse going through a 13 m highly nonlinear fiber (HNLF). The BiLSTM performed an RMSE metric of 0.004 and 0.012 for temporal and spectral domains, respectively. A supercontinuum generation was also performed in a 20-cm photonic-crystal fiber, and the BiLSTM returns an RMSE metric of 0.018 and 0.013 for temporal and spectral domains, respectively.

Improved monthly runoff time series prediction using the CABES-LSTM mixture model based on CEEMDAN-VMD decomposition

Abstract Accurate runoff prediction is vital in efficiently managing water resources. In this paper, a hybrid prediction model combining complete ensemble empirical mode decomposition with adaptive noise, variational mode decomposition, CABES, and long short-term memory network (CEEMDAN-VMD-CABES-LSTM) is proposed. Firstly, CEEMDAN is used to decompose the original data, and the high-frequency component is decomposed using VMD. Then, each component is input into the LSTM optimized by CABES for prediction. Finally, the results of individual component predictions are combined and reconstructed to produce the monthly runoff predictions. The hybrid model is employed to predict the monthly runoff at the Xiajiang hydrological station and the Yingluoxia hydrological station. A comprehensive comparison is conducted with other models including back propagation (BP), LSTM, etc. The assessment of each model's prediction performance uses four evaluation indexes. Results reveal that the CEEMDAN-VMD-CABES-LSTM model showcased the highest forecast accuracy among all the models evaluated. Compared with the single LSTM, the root mean square error (RMSE) and mean absolute percentage error (MAPE) of the Xiajiang hydrological station decreased by 71.09 and 65.26%, respectively, and the RMSE and MAPE of the Yingluoxia hydrological station decreased by 65.13 and 40.42%, respectively. The R and NSEC of both sites are near 1.

Research on Medium and Long Term Generation Side Deviation Prediction of New Power Market Based on Multi-Layer LSTM

Background: With the large-scale grid connection operation of new or renewable energy and the access of active loads such as electric vehicles and air conditioners, the electric energy trading business in the power market faces problems such as the rapid expansion of the number of market settlement subjects, explosive growth, various subjects responsible for deviation assessment, various electric energy trading methods and so on. Objective: This paper focuses on the medium and long-term generation side power trading in the new power market. Through cause analysis, induction and summary, algorithm design and case analysis, the problem of generation side deviation prediction is solved and power waste is reduced. Methods: This paper puts forward the reasons for the imbalance of medium and long-term power trading in the new power market dominated by new energy, as well as the deviation prediction algorithm based on multi-layer LSTM, which brings the total historical deviation, total planned deviation, total measurement deviation, new energy consumption and other data into the M-LSTM deep learning network for testing in each provincial power market center. Results: We use the neural network prediction algorithm. Compared with a single LSTM, the multi-layer LSTM can better maintain the characteristics of the sample time series and reduce the prediction error. Compared with BPNN、M-BPNN and Cooperative game theory, LSTM has a better memory effect. Conclusion: The experiment shows that the more accurate prediction deviation of this method can better arrange the generation plan, reduce the loss caused by excessive deviation, reduce the "price trampling" of the power market, and ensure the fair and efficient development of the power market.

Deep Learning in the Recognition of Activities of Daily Living Using Smartwatch Data.

The recognition of human activities (HAR) using wearable device data, such as smartwatches, has gained significant attention in the field of computer science due to its potential to provide insights into individuals' daily activities. This article aims to conduct a comparative study of deep learning techniques for recognizing activities of daily living (ADL). A mapping of HAR techniques was performed, and three techniques were selected for evaluation, along with a dataset. Experiments were conducted using the selected techniques to assess their performance in ADL recognition, employing standardized evaluation metrics, such as accuracy, precision, recall, and F1-score. Among the evaluated techniques, the DeepConvLSTM architecture, consisting of recurrent convolutional layers and a single LSTM layer, achieved the most promising results. These findings suggest that software applications utilizing this architecture can assist smartwatch users in understanding their movement routines more quickly and accurately.

A Stock Price Prediction Method based on LSTM and K-Means

Artificial neural network, as a nonlinear mapping or adaptive power system made by linking neurons,which can effectively resolve problems such as gradient explosion in stock price prediction processing. Recurrent neural network (RNN) is a common model for processing stock time-series data and is suitable for stock data involving sequential machine learning tasks, but the prediction results are poor when using long time span or nonlinear data for prediction. To address the problems of low prediction accuracy of ordinary neural networks for stock data with poor linearity and the inability of a single LSTM model to show the recommendation level of a target stock, the paper proposes a deep learning factor integrated prediction model based on LSTM-K-Means.On this basis, a stock price prediction method based on long-term and short-term memory network namely LSTM and K-means clustering algorithm is proposed. The method is not only designed to model stock ups and downs at different levels of combinations, but also more intuitively identifies stocks with better ups through returns and volatilities. Through experimental verification, the stock price prediction method based on LSTM with K-Means proposed in this paper is effective.

Spatio-temporal prediction for distributed PV generation system based on deep learning neural network model

To obtain higher accuracy of PV prediction to enhance PV power generation technology. This paper proposes a spatio-temporal prediction method based on a deep learning neural network model. Firstly, spatio-temporal correlation analysis is performed for 17 PV sites. Secondly, we compare CNN-LSTM with a single CNN or LSTM model trained on the same dataset. From the evaluation indexes such as loss map, regression map, RMSE, and MAE, the CNN-LSTM model that considers the strong correlation of spatio-temporal correlation among the 17 sites has better performance. The results show that our method has higher prediction accuracy.