Long-term Temporal Dependencies Research Articles

A dual-model deep learning method for sleep apnea detection based on representation learning and temporal dependence

Sleep apnea (SA) is a sleep-breathing disorder accompanied by multiple complications. The SA detection method based on a single-lead electrocardiogram (ECG) has the characteristics of low power consumption and is desirable for the development of wearable equipment. This study proposed a dual-model deep learning method to perform representation learning and introduce long-term temporal dependence. First, the Christov algorithm was used to obtain the RR interval (RRI) of each 1-minute ECG segment, and the adaptive synthetic (ADASYN) sampling method was employed to synthesize the RRI series of the minority class to address the imbalanced learning problem. Then, a representation learning model based on the one-dimensional convolutional neural network (1DCNN-RLM) was built to extract the feature vector of the RRI series. Eventually, a temporal dependence model based on the bidirectional gated recurrent unit (BiGRU-TDM) was constructed to learn the state (SA/normal) transition pattern between the segments and complete the classification task. We employed the apnea-ECG database for experiments. For per-segment detection results, the accuracy, sensitivity, and specificity of this method were 91.1%, 88.9%, and 92.4%, respectively. The per-recording detection accuracy reached 100%. ADASYN alleviates the imbalance of sensitivity and specificity in classification results. The 1DCNN-RLM with powerful representation learning ability has extracted discriminative features. The BiGRU-TDM introduces the long-term time dependence of SA and improves classification performance. The results of this study substantiate that the proposed method is robust and has good transferability. This method provides a reference for the diagnosis of other diseases.

SA-LSTMs: A new advance prediction method of energy consumption in cement raw materials grinding system

Electricity consumption is a major energy efficiency indicator in cement raw materials grinding system. Advance prediction of electricity consumption provides the basis for cement production scheduling and achieves the energy saving. However, due to the influence of strong coupling, delay, intrinsic non-linearity and uncertainty, it is difficult to model the raw materials grinding system accurately. To address this problem, a new model of long short-term memory networks based on spatial attention is proposed in this paper. The proposed method uses LSTM neural networks appropriately to capture the long-term temporal dependencies of the system variables appropriately to solve the problem of time delay. Spatial attention is designed to improve the spatial perception of the model and increase the sensitivity to important information. Thus, the proposed method can efficiently capture the temporal and spatial characteristics of the cement raw materials grinding system. In addition, to verify the superiority of the proposed method, we introduce LSTM, Seq2Seq, ARIMA, SVM and XGBoost for comparison. The experimental result confirms that the proposed model achieves the highest performance, which provides an encouraging perspective for real-time energy consumption prediction in raw materials grinding.

Epidemiological Predictive Modeling of COVID-19 Infection: Development, Testing, and Implementation on the Population of the Benelux Union.

Since the outbreak of coronavirus disease-2019 (COVID-19), the whole world has taken interest in the mechanisms of its spread and development. Mathematical models have been valuable instruments for the study of the spread and control of infectious diseases. For that purpose, we propose a two-way approach in modeling COVID-19 spread: a susceptible, exposed, infected, recovered, deceased (SEIRD) model based on differential equations and a long short-term memory (LSTM) deep learning model. The SEIRD model is a compartmental epidemiological model with included components: susceptible, exposed, infected, recovered, deceased. In the case of the SEIRD model, official statistical data available online for countries of Belgium, Netherlands, and Luxembourg (Benelux) in the period of March 15 2020 to March 15 2021 were used. Based on them, we have calculated key parameters and forward them to the epidemiological model, which will predict the number of infected, deceased, and recovered people. Results show that the SEIRD model is able to accurately predict several peaks for all the three countries of interest, with very small root mean square error (RMSE), except for the mild cases (maximum RMSE was 240.79 ± 90.556), which can be explained by the fact that no official data were available for mild cases, but this number was derived from other statistics. On the other hand, LSTM represents a special kind of recurrent neural network structure that can comparatively learn long-term temporal dependencies. Results show that LSTM is capable of predicting several peaks based on the position of previous peaks with low values of RMSE. Higher values of RMSE are observed in the number of infected cases in Belgium (RMSE was 535.93) and Netherlands (RMSE was 434.28), and are expected because of thousands of people getting infected per day in those countries. In future studies, we will extend the models to include mobility information, variants of concern, as well as a medical intervention, etc. A prognostic model could help us predict epidemic peaks. In that way, we could react in a timely manner by introducing new or tightening existing measures before the health system is overloaded.

DeepSense: A Physics-Guided Deep Learning Paradigm for Anomaly Detection in Soil Gas Data at Geologic CO2 Storage Sites.

Driven by the collection of enormous amounts of streaming data from sensors, and with the emergence of the internet of things, the need for developing robust detection techniques to identify data anomalies has increased recently. The algorithms for anomaly detection are required to be selected based on the type of data. In this study, we propose a predictive anomaly detection technique, DeepSense, which is applied to soil gas concentration data acquired from sensors being used for environmental characterization at a prospective CO2 storage site in Queensland, Australia. DeepSense takes advantage of deep-learning algorithms as its predictor module and uses a process-based soil gas method as the basis of its anomaly detector module. The proposed predictor framework leverages the power of convolutional neural network algorithms for feature extraction and simultaneously captures the long-term temporal dependency through long short-term memory algorithms. The proposed process-based anomaly detection method is a cost-effective alternative to the conventional concentration-based soil gas methodologies which rely on long-term baseline surveys for defining the threshold level. The results indicate that the proposed framework performs well in diagnosing anomalous data in soil gas concentration data streams. The robustness and efficacy of the DeepSense were verified against data sets acquired from different monitoring stations of the storage site.

Spatial-Temporal Feature-Based Sports Video Classification

Video classification has been an active research field of computer vision in last few years. Its main purpose is to produce a label that is relevant to the video given its frames. Unlike image classification, which takes still pictures as input, the input of video classification is a sequence of images. The complex spatial and temporal structures of video sequence incur understanding and computation difficulties, which should be modeled to improve the video classification performance. This work focuses on sports video classification but can be expanded into other applications. In this paper, the authors propose a novel sports video classification method by processing the video data using convolutional neural network (CNN) with spatial attention mechanism and deep bidirectional long short-term memory (BiLSTM) network with temporal attention mechanism. The method first extracts 28 frames from each input video and uses the classical pre-trained CNN to extract deep features, and the spatial attention mechanism is applied to CNN features to decide ‘where' to look. Then the BiLSTM is utilized to model the long-term temporal dependence between video frame sequences, and the temporal attention mechasim is employed to decide ‘when' to look. Finally, the label of the input video is given by the classification network. In order to evaluate the feasibility and effectiveness of the proposed method, an extensive experimental investigation was conducted on the open challenging sports video datasets of Sports8 and Olympic16; the results show that the proposed CNN-BiLSTM network with spatial temporal attention mechanism can effectively model the spatial-temporal characteristics of video sequences. The average classification accuracy of the Sports8 is 98.8%, which is 6.8% higher than the existing method. The average classification accuracy of 90.46% is achieved on Olympic16, which is about 18% higher than the existing methods. The performance of the proposed approach outperforms the state-of-the-art methods, and the experimental results demonstrate the effectiveness of the proposed approach.

Multi-view deep forecasting for hourly solar irradiance with error correction

Short-term solar irradiance forecasting is crucial in managing power network operations and solar photovoltaic applications. In this paper, a Multi-view Deep Forecasting method with Error Correction (MvDF_EC) for 1-hour ahead solar forecasting is proposed. MvDF_EC comprises of the Multi-view Deep Forecasting method (MvDF) and a robust Radial Basis Function Neural Network trained via minimizing the Localized Generalization Error for compensating the solar forecasting error of MvDF. MvDF consists of three deep neural networks which learn representations of input data from different views. The three views are 1) the hierarchical local temporal information extracted by the Temporal Convolutional Neural Network (TCN), 2) the key context sequential information captured by the Bi-directional Long Short-Term Memory Neural Network with Temporal Attention (BLSTMattn), and 3) long-term temporal dependencies between local temporal patterns filtered by the Convolutional Gated Recurrent Unit Neural Network (C_GRU). The solar forecasting performance of the proposed MvDF_EC is evaluated with the National Solar Radiation Database. Simulation results show that MvDF_EC yields the most accurate solar prediction compared with the benchmarks including the smart persistence and the state-of-the-art models. The lowest relative Root Mean Square Error values for Maraba and Labelle are 22.08% and 27.40%, respectively in 1-hour ahead solar forecasting.

A novel deep learning prediction model for concrete dam displacements using interpretable mixed attention mechanism

Dam displacements can effectively reflect its operational status, and thus establishing a reliable displacement prediction model is important for dam health monitoring. The majority of the existing data-driven models, however, focus on static regression relationships, which cannot capture the long-term temporal dependencies and adaptively select the most relevant influencing factors to perform predictions. Moreover, the emerging modeling tools such as machine learning (ML) and deep learning (DL) are mostly black-box models, which makes their physical interpretation challenging and greatly limits their practical engineering applications. To address these issues, this paper proposes an interpretable mixed attention mechanism long short-term memory (MAM-LSTM) model based on an encoder-decoder architecture, which is formulated in two stages. In the encoder stage, a factor attention mechanism is developed to adaptively select the highly influential factors at each time step by referring to the previous hidden state. In the decoder stage, a temporal attention mechanism is introduced to properly extract the key time segments by identifying the relevant hidden states across all the time steps. For interpretation purpose, our emphasis is placed on the quantification and visualization of factor and temporal attention weights. Finally, the effectiveness of the proposed model is verified using monitoring data collected from a real-world dam, where its accuracy is compared to a classical statistical model, conventional ML models, and homogeneous DL models. The comparison demonstrates that the MAM-LSTM model outperforms the other models in most cases. Furthermore, the interpretation of global attention weights confirms the physical rationality of our attention-based model. This work addresses the research gap in interpretable artificial intelligence for dam displacement prediction and delivers a model with both high-accuracy and interpretability.

Bidirectional LSTM with saliency-aware 3D-CNN features for human action recognition

Deep convolutional neural network (DCNN) and recurrent neural network (RNN) have been proved as an imperious research area in multimedia understanding and obtained remarkable action recognition performance. However, videos contain rich motion information with varying dimensions. Existing recurrent based pipelines fail to capture long-term motion dynamics in videos with various motion scales and complex actions performed by multiple actors. Consideration of contextual and salient features is more important than mapping a video frame into a static video representation. This research work provides a novel pipeline by analyzing and processing the video information using a 3D convolution (C3D) network and newly introduced deep bidirectional LSTM. Like popular two-stream convent, we also introduce a two-stream framework with one modification; that is, we replace the optical flow stream by saliency-aware stream to avoid the computational complexity. First, we generate a saliency-aware video stream by applying the saliency-aware method. Secondly, a two-stream 3D-convolutional network (C3D) is utilized with two different types of streams, i.e., RGB stream and saliency-aware video stream, to collect both spatial and semantic temporal features. Next, a deep bidirectional LSTM network is used to learn sequential deep temporal dynamics. Finally, time-series-pooling-layer and softmax-layers classify human activity and behavior. The introduced system can learn long-term temporal dependencies and can predict complex human actions. Experimental results demonstrate the significant improvement in action recognition accuracy on different benchmark datasets.

Time-Wise Attention Aided Convolutional Neural Network for Data-Driven Cellular Traffic Prediction

Recurrent neural network (RNN) based models are widely adopted to capture temporal dependencies in the state-of-the-art approaches for cellular traffic prediction. However, RNN is inefficient and incapable of capturing long-range temporal dependencies of traffic data. Besides, its inherent sequential nature makes it time consuming in capture the temporal dependencies. To better capture the long-term temporal dependency and reduce the consumed time in traffic data prediction, we propose a time-wise attention aided convolutional neural network (TWACNet) structure for cellular traffic prediction. In the proposed TWACNet, the time-wise attention mechanism is adopted to capture long-range temporal dependencies of the cellular traffic data and the convolutional neural network (CNN) is adopted to capture the spatial correlation. The performance of TWACNet in traffic prediction is tested in real-world cellular traffic datasets. Experimental results demonstrate that our proposed approach can considerably outperform those existing prediction methods in terms of root mean square errors (RMSE) and training time.

Deep probabilistic tracking of particles in fluorescence microscopy images.

Tracking of particles in temporal fluorescence microscopy image sequences is of fundamental importance to quantify dynamic processes of intracellular structures as well as virus structures. We introduce a probabilistic deep learning approach for fluorescent particle tracking, which is based on a recurrent neural network that mimics classical Bayesian filtering. Compared to previous deep learning methods for particle tracking, our approach takes into account uncertainty, both aleatoric and epistemic uncertainty. Thus, information about the reliability of the computed trajectories is determined. Manual tuning of tracking parameters is not necessary and prior knowledge about the noise statistics is not required. Short and long-term temporal dependencies of individual object dynamics are exploited for state prediction, and assigned detections are used to update the predicted states. For correspondence finding, we introduce a neural network which computes assignment probabilities jointly across multiple detections as well as determines the probabilities of missing detections. Training requires only simulated data and therefore tedious manual annotation of ground truth is not needed. We performed a quantitative performance evaluation based on synthetic and real 2D as well as 3D fluorescence microscopy images. We used image data of the Particle Tracking Challenge as well as real time-lapse fluorescence microscopy images displaying virus structures and chromatin structures. It turned out that our approach yields state-of-the-art results or improves the tracking results compared to previous methods.

CPMAN: Change Point Detection Approach in Time Series Based on the Prediction of Multi-stage Attention Networks

Time series change point detection can identify the locations of abrupt points in many dynamic processes. It can help us to find anomaly data in an early stage. At the same time, detecting change points for long, periodic, and multiple input series data has received a lot of attention recently, and is universally applicable in many fields including power, environment, finance, and medicine. However, the performance of classical methods typically scales poorly for such time series. In this paper, we propose CPMAN, a novel prediction-based change point detection approach via multi-stage attention networks. Our model consists of two key modules. First, in the time series prediction module, we employ the multi-stage attention-based networks and integrate the multi-series fusion mechanism. This module can adaptively extract features from the relevant input series and capture the long-term temporal dependencies. Secondly, in the change point detection module, we use the wavelet analysis-based algorithm to detect change points efficiently and identify the change points and outliers. Extensive experiments are conducted on various real-world datasets and synthetic datasets, proving the superiority and effectiveness of CPMAN. Our approach outperforms the state-of-the-art methods by up to 12.1% on the F1 Score.

End-to-end speech emotion recognition using a novel context-stacking dilated convolution neural network

Amongst the various characteristics of a speech signal, the expression of emotion is one of the characteristics that exhibits the slowest temporal dynamics. Hence, a performant speech emotion recognition (SER) system requires a predictive model that is capable of learning sufficiently long temporal dependencies in the analysed speech signal. Therefore, in this work, we propose a novel end-to-end neural network architecture based on the concept of dilated causal convolution with context stacking. Firstly, the proposed model consists only of parallelisable layers and is hence suitable for parallel processing, while avoiding the inherent lack of parallelisability occurring with recurrent neural network (RNN) layers. Secondly, the design of a dedicated dilated causal convolution block allows the model to have a receptive field as large as the input sequence length, while maintaining a reasonably low computational cost. Thirdly, by introducing a context stacking structure, the proposed model is capable of exploiting long-term temporal dependencies hence providing an alternative to the use of RNN layers. We evaluate the proposed model in SER regression and classification tasks and provide a comparison with a state-of-the-art end-to-end SER model. Experimental results indicate that the proposed model requires only 1/3 of the number of model parameters used in the state-of-the-art model, while also significantly improving SER performance. Further experiments are reported to understand the impact of using various types of input representations (i.e. raw audio samples vs log mel-spectrograms) and to illustrate the benefits of an end-to-end approach over the use of hand-crafted audio features. Moreover, we show that the proposed model can efficiently learn intermediate embeddings preserving speech emotion information.

Automatic and non-invasive Parkinson’s disease diagnosis and severity rating using LSTM network

Deep learning has a huge potential in healthcare for uncovering the hidden patterns from large volume of clinical data to diagnose different diseases. This paper presents a novel deep learning architecture based long short term memory (LSTM) network for severity rating of Parkinson’s disease (PD) using gait pattern. Unlike machine learning (ML) algorithms, the LSTM network avoids the need for hand crafted features and learns the long-term temporal dependencies in the gait cycle for robust diagnosis of PD. The primary advantage of the LSTM network is that it solves the vanishing gradient problem by introducing the memory blocks in place of self-connected hidden units, thereby deciding when to learn new information. Three distinct gait datasets containing vertical ground reaction force (VGRF) recordings for different walking scenarios are used for training the LSTM network. To avoid data overfitting, the proposed approach utilizes dropout and L2 regularization techniques. For solving the cost function, Adam, a stochastic gradient-based optimizer, is employed and the severity of PD is categorized based on unified Parkinson’s disease rating scale (UPDRS) and Hoehn and Yahr (H&Y) scale. The experimental results reveal that Adam optimized LSTM network can effectively learn the gait kinematic features and offer an average accuracy of 98.6% for binary classification and 96.6% for multi-class classification, with an accuracy improvement of 3.4% in comparison with the related techniques.

A Long-term Recurrent Convolutional Network for Stock Index Prediction

Stock index prediction aims to predict the future price of stock indexes, which plays a key role in seeking the maximum profit from stock investment. However, It has been proven to be a very difficult task because of its innate complexity, dynamics, and uncertainty. With the rapid development of deep learning, more researchers have attempted to apply nonlinear learning methods such as long short-term memory networks (LSTMs) to capture the complex patterns hidden in market trends. In this paper, we propose a Long-term Recurrent Convolutional Network (LRCN), which combines convolutional layers and long-range temporal recursion and is end-to-end trainable. In the LRCN model, the two-dimensional convolutional neural network (2D-CNN) performs convolution on the most recent region to capture local fluctuation features, and the long short-term memory (LSTM) learns the long-term temporal dependencies to improve stock index prediction. To evaluate the effectiveness of LRCN, we collected real stock market data for stock indexes S&P 500 and DJIA, and the experimental results show that the proposed LRCN can significantly outperform several existing highly competitive methods.

Important Trading Point Prediction Using a Hybrid Convolutional Recurrent Neural Network

Stock performance prediction plays an important role in determining the appropriate timing of buying or selling a stock in the development of a trading system. However, precise stock price prediction is challenging because of the complexity of the internal structure of the stock price system and the diversity of external factors. Although research on forecasting stock prices has been conducted continuously, there are few examples of the successful use of stock price forecasting models to develop effective trading systems. Inspired by the process of human stock traders looking for trading opportunities, we propose a deep learning framework based on a hybrid convolutional recurrent neural network (HCRNN) to predict the important trading points (IPs) that are more likely to be followed by a significant stock price rise to capture potential high-margin opportunities. In the HCRNN model, the convolutional neural network (CNN) performs convolution on the most recent region to capture local fluctuation features, and the long short-term memory (LSTM) approach learns the long-term temporal dependencies to improve stock performance prediction. Comprehensive experiments on real stock market data prove the effectiveness of our proposed framework. Our proposed method ITPP-HCRNN achieves an annualized return that is 278.46% more than that of the market.

Combining a parallel 2D CNN with a self-attention Dilated Residual Network for CTC-based discrete speech emotion recognition

A challenging issue in the field of the automatic recognition of emotion from speech is the efficient modelling of long temporal contexts. Moreover, when incorporating long-term temporal dependencies between features, recurrent neural network (RNN) architectures are typically employed by default. In this work, we aim to present an efficient deep neural network architecture incorporating Connectionist Temporal Classification (CTC) loss for discrete speech emotion recognition (SER). Moreover, we also demonstrate the existence of further opportunities to improve SER performance by exploiting the properties of convolutional neural networks (CNNs) when modelling contextual information. Our proposed model uses parallel convolutional layers (PCN) integrated with Squeeze-and-Excitation Network (SEnet), a system herein denoted as PCNSE, to extract relationships from 3D spectrograms across timesteps and frequencies; here, we use the log-Mel spectrogram with deltas and delta–deltas as input. In addition, a self-attention Residual Dilated Network (SADRN) with CTC is employed as a classification block for SER. To the best of the authors’ knowledge, this is the first time that such a hybrid architecture has been employed for discrete SER. We further demonstrate the effectiveness of our proposed approach on the Interactive Emotional Dyadic Motion Capture (IEMOCAP) and FAU-Aibo Emotion corpus (FAU-AEC). Our experimental results reveal that the proposed method is well-suited to the task of discrete SER, achieving a weighted accuracy (WA) of 73.1% and an unweighted accuracy (UA) of 66.3% on IEMOCAP, as well as a UA of 41.1% on the FAU-AEC dataset.

A Multiscale Interactive Recurrent Network for Time-Series Forecasting.

Time-series forecasting is a key component in the automation and optimization of intelligent applications. It is not a trivial task, as there are various short-term and/or long-term temporal dependencies. Multiscale modeling has been considered as a promising strategy to solve this problem. However, the existing multiscale models either apply an implicit way to model the temporal dependencies or ignore the interrelationships between multiscale subseries. In this article, we propose a multiscale interactive recurrent network (MiRNN) to jointly capture multiscale patterns. MiRNN employs a deep wavelet decomposition network to decompose the raw time series into multiscale subseries. MiRNN introduces three key strategies (truncation, initialization, and message passing) to model the inherent interrelationships between multiscale subseries, as well as a dual-stage attention mechanism to capture multiscale temporal dependencies. Experiments on four real-world datasets demonstrate that our model achieves promising performance compared with the state-of-the-art methods.

Spatial-Temporal Action Localization With Hierarchical Self-Attention

This paper proposes a novel architecture for spatial-temporal action localization in videos. The new architecture first employs a two-stream 3D convolutional neural network (3D-CNN) to provide initial action detection. Next, a new Hierarchical Self-Attention Network (HiSAN), the core of this architecture, is developed to learn the spatial-temporal relationships of key actors. Spatial Gaussian priors (SGP) are also imbued to the bidirectional self-attention to enhance HiSAN in modelling the relationships of neighboring actors. Such a combination of 3D-CNN and SGP augmented HiSAN allows us to effectively extract both of the spatial context information and the long-term temporal dependency to improve action localization accuracy. Afterwards, a new fusion strategy is employed, which first re-scores the bounding boxes to settle the inconsistent detection scores caused by background clutter or occlusion, and then aggregates the motion and appearance information from the two-stream network with the motion saliency to alleviate the impact of camera movement. Finally, a tube association network based on the self-similarity of the actors’ appearance and spatial information across frames is addressed to efficaciously construct the action tubes. Simulations on four widespread datasets reveal the efficacy of the new approach.

PatientFlowNet: A Deep Learning Approach to Patient Flow Prediction in Emergency Departments

Emergency Department (ED) crowding is a major public health challenge since it can seriously impact patient outcomes; and accurate prediction of patient flow in EDs is essential for improving operational efficiency and quality of care. We present a deep learning framework to predict patient flow rates in EDs, namely the rates of arrival, treatment, and discharge for patients across all triage levels. Our model detects short-term and long-term temporal dependencies within the time-series data of a given patient-flow variable, as well as dependencies between time-series data of different patient-flow variables. We implement this framework as a convolutional neural network, which we call PatientFlowNet. Our proposed model learns simultaneously from multiple flow variables over a long temporal window and predicts future values of arrival, treatment, and discharge rates in the ED. We benchmark our model against state-of-the-art methods on data from EDs in three different hospitals. Results show that PatientFlowNet achieves superior prediction accuracy, compared to the baseline methods, and yields a mean absolute error that is 4.8% lower than the leading baseline. Furthermore, we provide a visual and interpretable representation of the learned dependencies by our model, between patient-flow variables in EDs.

Electricity Consumption Forecasting Using Gated-FCN With Ensemble Strategy

Accurate electricity consumption forecasting in the power grids ensures efficient generation and distribution of electricity. Keeping this in mind, the paper introduces a novel deep learning model, termed Gated-FCN, for short-term load forecasting. The key idea is to introduce an automated feature selection and deep learning model for forecasting. The model includes an eight-layered Fully Convolutional Network (FCN-8) in which the hand-crafted feature selection that requires expert domain knowledge is avoided. Furthermore, the model also reduces noise as it learns internal dependencies and the correlation of the time series. Enhanced Bidirectional Gated Recurrent Unit (EBiGRU) is used in combination with FCN-8 to learn long-term temporal dependencies of the time series. Moreover, a weighted averaging mechanism of multiple snapshot models is adopted in the proposed model to assign optimized weights to BiGRU. At the end of FCN-8 and BiGRU, a fully connected dense layer is used that gives final prediction results. Gated-FCN is an end-to-end forecasting model that does not require any other model for enhancing its forecasting efficiency. Different activation functions are initially analyzed to determine how the proposed model learns complex patterns from the time series data. Later, the activation function having the best accuracy is used for forecasting. The proposed model extracts both spatial and temporal features from the data. Furthermore, this paper also provides predictive and exploratory data analyses to assist policymakers in making optimal decisions regarding power production and dispatch. In order to demonstrate the applicability of the proposed technique, the simulations are performed using nine years’ load consumption data taken from Independent System Operators New England (ISO-NE). The comparison with five state-of-the-art techniques is also provided to prove the fact that Gated-FCN gives the best forecasting accuracy as compared to other benchmark techniques in terms of two performance metrics: Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE).