Search

Driving risk prediction is crucial for advanced driving technologies, with deep learning approaches leading the way in driving safety analysis. Current driving risk prediction methods typically establish a mapping between driving features and risk statuses. However, status prediction fails to provide detailed risk sequence information, and existing driving safety analyses seldom focus on run-off-road (ROR) risk. This study extracted 660 near-roadside lane-changing samples from the high-D natural driving dataset. The performance of sequence and status prediction for ROR risk was compared across five mainstream deep learning models: LSTM, CNN, LSTM-CNN, CNN-LSTM-MA, and Transformer. The results indicate the following: (1) The deep learning approach effectively predicts ROR risk. The Macro F1 Score of sequence prediction significantly surpasses that of status prediction, with no notable difference in efficiency; (2) Sequence prediction captures risk evolution trends, such as increases, turns, and declines, providing more comprehensive safety information; (3) The presence of surrounding vehicles significantly impacts lane change duration and ROR risk. This study offers new insights into the quantitative research of ROR risk, demonstrating that risk sequence prediction is superior to status prediction in multiple aspects and can provide theoretical support for the development of roadside safety. [ABSTRACT FROM AUTHOR]

The deep learning steel plastic constitutive model training method was studied based on the recurrent neural network (RNN) model to improve the allocative efficiency of the deep learning steel plastic constitutive model and promote its application in practical engineering. Two linear hardening constitutive datasets of steel were constructed using the Gaussian stochastic process. The RNN, long short-term memory (LSTM), and gated recurrent unit (GRU) were used as models for training. The effects of the data pre-processing method, neural network structure, and training method on the model training were analyzed. The prediction ability of the model for different scale series and the corresponding data demand were evaluated. The results show that LSTM and the GRU are more suitable for stress–strain prediction. The marginal effect of the stacked neural network depth and number gradually decreases, and the hysteresis curve can be accurately predicted by a two-layer RNN. The optimal structure of the two models is A50-100 and B150-150. The prediction accuracy of the models increased with the decrease in batch size and the increase in training batch, and the training time also increased significantly. The decay learning rate method could balance the prediction accuracy and training time, and the optimal initial learning rate, batch size, and training batch were 0.001, 60, and 100, respectively. The deep learning plastic constitutive model based on the optimal parameters can accurately predict the hysteresis curve of steel, and the prediction abilities of the GRU are 6.13, 6.7, and 3.3 times those of LSTM in short, medium, and long sequences, respectively. [ABSTRACT FROM AUTHOR]

Recently, significant developments have been made in Long Short-Term Memory (LSTM) networks within the realm of synthesis music. Notwithstanding these advancements, several challenges persist warranting further research. Primarily, there exists an absence of dedicated research on the application of LSTM networks for the synthesis of Original Sound Tracks (OST). Secondly, in general, people can only judge whether the synthesized music meets their expectations based on the model output. However, due to the time-consuming of training the model may need to try multiple times to obtain successful training results. Moreover, the subjective of music quality evaluation relying on human perception, not only the result of model training. To address these multifaceted challenges, this paper concentrates specifically on OST and proposes a framework termed the OST Synthesis Framework (OSTSF) utilizing LSTM. This framework accepts various OST types as input, processed through LSTM to yield innovative OST. Additionally, a novel preprocessing algorithm is proposed to screen input OST elements such as notes and chords, enabling control over music type and quality before the training phase. This algorithm serves to mitigate training uncertainties and reduce situations that require repeated training. Besides, a postprocessing approach, leveraging mathematical formulations facilitates the evaluation of synthesis OST also proposed. This approach aims to quantify subjective evaluations, providing a more intuitive representation through scoring metrics. Experiment results reveal that the OSTSF synthesized OST received favorable rate among a cohort of 100 surveyed respondents attaining 78.8%, demonstrating the efficacy of the proposed framework in the realm of music synthesis utilizing LSTM.

Driving risk prediction is crucial for advanced driving technologies, with deep learning approaches leading the way in driving safety analysis. Current driving risk prediction methods typically establish a mapping between driving features and risk statuses. However, status prediction fails to provide detailed risk sequence information, and existing driving safety analyses seldom focus on run-off-road (ROR) risk. This study extracted 660 near-roadside lane-changing samples from the high-D natural driving dataset. The performance of sequence and status prediction for ROR risk was compared across five mainstream deep learning models: LSTM, CNN, LSTM-CNN, CNN-LSTM-MA, and Transformer. The results indicate the following: (1) The deep learning approach effectively predicts ROR risk. The Macro F1 Score of sequence prediction significantly surpasses that of status prediction, with no notable difference in efficiency; (2) Sequence prediction captures risk evolution trends, such as increases, turns, and declines, providing more comprehensive safety information; (3) The presence of surrounding vehicles significantly impacts lane change duration and ROR risk. This study offers new insights into the quantitative research of ROR risk, demonstrating that risk sequence prediction is superior to status prediction in multiple aspects and can provide theoretical support for the development of roadside safety.

The deep learning steel plastic constitutive model training method was studied based on the recurrent neural network (RNN) model to improve the allocative efficiency of the deep learning steel plastic constitutive model and promote its application in practical engineering. Two linear hardening constitutive datasets of steel were constructed using the Gaussian stochastic process. The RNN, long short-term memory (LSTM), and gated recurrent unit (GRU) were used as models for training. The effects of the data pre-processing method, neural network structure, and training method on the model training were analyzed. The prediction ability of the model for different scale series and the corresponding data demand were evaluated. The results show that LSTM and the GRU are more suitable for stress–strain prediction. The marginal effect of the stacked neural network depth and number gradually decreases, and the hysteresis curve can be accurately predicted by a two-layer RNN. The optimal structure of the two models is A50-100 and B150-150. The prediction accuracy of the models increased with the decrease in batch size and the increase in training batch, and the training time also increased significantly. The decay learning rate method could balance the prediction accuracy and training time, and the optimal initial learning rate, batch size, and training batch were 0.001, 60, and 100, respectively. The deep learning plastic constitutive model based on the optimal parameters can accurately predict the hysteresis curve of steel, and the prediction abilities of the GRU are 6.13, 6.7, and 3.3 times those of LSTM in short, medium, and long sequences, respectively.

A key problem domain inside Robotic Process Automation is the automatic discovery of workflow process schemes. Considering current process mining technologies, graph-based approaches dominate the industry. On the other hand, the conventional methods suffer from low time efficiency and varying accuracy. Machine learning-based methods can provide better efficiency, but they have significant limitations considering schema flexibility. The paper presents a novel neural network-based schema induction model for the discovery of event patterns containing parallel and optional sequences of different actors. This model can process more complex event graphs and situations than the conventional methods. The performed analysis and test results show the unique power of this approach in process schema mining. [ABSTRACT FROM AUTHOR]

Abstract Apart from nodes and links, for many networked systems, we have access to data on paths, i.e., collections of temporally ordered variable-length node sequences that are constrained by the system’s topology. Understanding the patterns in such data is key to advancing our understanding of the structure and dynamics of complex systems. Moreover, the ability to accurately model and predict paths is important for engineered systems, e.g., to optimise supply chains or provide smart mobility services. Here, we introduce MOGen, a generative modelling framework that enables both next-element and out-of-sample prediction in paths with high accuracy and consistency. It features a model selection approach that automatically determines the optimal model directly from data, effectively making MOGen parameter-free. Using empirical data, we show that our method outperforms state-of-the-art sequence modelling techniques. We further introduce a mathematical formalism that links higher-order models of paths to transition matrices of random walks in multi-layer networks.

Background: The relationship between the sequence of a protein, its structure, and the resulting connection between its structure and function, is a foundational principle in biological science. Only recently has the computational prediction of protein structure based only on protein sequence been addressed effectively by AlphaFold, a neural network approach that can predict the majority of protein structures with X-ray crystallographic accuracy. A question that is now of acute relevance is the "inverse protein folding problem": predicting the sequence of a protein that folds into a specified structure. This will be of immense value in protein engineering and biotechnology, and will allow the design and expression of recombinant proteins that can, for instance, fold into specified structures as a scaffold for the attachment of recombinant antigens, or enzymes with modified or novel catalytic activities. Here we describe the development of SeqPredNN, a feed-forward neural network trained with X-ray crystallographic structures from the RCSB Protein Data Bank to predict the identity of amino acids in a protein structure using only the relative positions, orientations, and backbone dihedral angles of nearby residues. Results: We predict the sequence of a protein expected to fold into a specified structure and assess the accuracy of the prediction using both AlphaFold and RoseTTAFold to computationally generate the fold of the derived sequence. We show that the sequences predicted by SeqPredNN fold into a structure with a median TM-score of 0.638 when compared to the crystal structure according to AlphaFold predictions, yet these sequences are unique and only 28.4% identical to the sequence of the crystallized protein. Conclusions: We propose that SeqPredNN will be a valuable tool to generate proteins of defined structure for the design of novel biomaterials, pharmaceuticals, catalysts, and reporter systems. The low sequence identity of its predictions compared to the native sequence could prove useful for developing proteins with modified physical properties, such as water solubility and thermal stability. The speed and ease of use of SeqPredNN offers a significant advantage over physics-based protein design methods. [ABSTRACT FROM AUTHOR]

Dementia is not a specific disease, but a general term for age-related decline or loss of memory, cognitive abilities including problem solving and decision-making, and one’s own language, which significantly interfere with daily life. Researchers around the world have developed ways to automate the diagnosis of dementia through the use of machine learning and data mining approaches. The aim of this research project is to design and develop a day-to-day activity prediction algorithm in order to accurately identify and differentiate the dementia affected patients from the healthy subjects, to ensure early diagnosis of dementia development. This research advocates a novel algorithm called ‘Sequence Prediction via All Discoverable Episodes (SPADE)’ as a statistical tool to map activities of daily life (ADLs) in different groups of people in order to develop a unique parameter for precise diagnosis. The results of our experiment demonstrated a significant difference (i.e. 11 %) in the sequence prediction peak accuracy between the healthy subjects and the residents with dementia. SPADE demonstrated an adequate accuracy (i.e. 80 % on average), with an improvement of about 12 % compared to the performance of M-SPEED in inferring future occurrences of activities. It is thus evident that the algorithms for activity predictions show promise for early detection of dementia symptoms without the use of any expensive clinical procedure.

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This article develops a recurrent neural network (RNN) with an encoder–decoder structure to predict the driving sequence of SiC MOSFET active gate drivers (AGDs). With a set of switching targets as the input, the predictor generates an optimal active gate driving sequence to improve the switching transient. The development is based on a hybrid platform across MATLAB, PyTorch, and LTspice. A high-fidelity switching model is implemented in MATLAB to obtain reliable training data. The sequence predictor is trained with PyTorch. The predicted sequence is verified on an example Buck circuit in LTspice. In contrast to the state-of-the-art approach, the proposed method avoids exhaustive search in a large solution space; the sequence length is dynamically predicted per the driving strength at each step. The AGD sequences generated by the predictor effectively and precisely improve the switching transients, making the proposed sequence predictor an integral and valuable component for active gate driving.

Satellite cloud images can help meteorologists characterize the weather patterns, such as identifying tropical cyclone (TC) intensity, detecting climate anomaly regions and predicting rain effects, which makes satellite cloud image forecasting become an important task. In recent years, an increasing number of deep learning models have demonstrated their ability to predict spatiotemporal types of data. However, applying these methods directly to predict satellite cloud images does not consider the chaotic nature of the atmosphere and the satellite images of large scale observation areas, especially in extreme climate events, such as tropical cyclones. Hence, we propose both a novel deep learning model called "Satellite Cloud Spatio-Temporal sequence" (SCSTque) and a dataset benchmark called "Tropical Cyclone Cloudage Map Dataset" (TCCMD) for tropical cyclone satellite images to address these problems. Specifically, the SCSTque model is a sequence-to-sequence (Seq2Seq) autoencoder architecture to use the previous images to predict the next few images, which can fully extract spatial appearance features and temporal dynamics. The TCCMD dataset is a real-world large-scale dataset, which contains satellite image sequences for each corresponding tropical cyclone event from 2010 to 2020. The novel method is validated using the TCCMD dataset and the Moving MNIST dataset. The experimental results proved the SCSTque method outperforms the baseline Seq2Seq-based methods, including a ConvLSTM model, a PredRNN model and a FCLSTM model. [ABSTRACT FROM AUTHOR]

Sixty percent of the failures of launch vehicles in the ascending phase occur in the propulsion system. Among them, the vibration generated by the engine is an important factor in the occurrence of failure. At present, health assessment methods in the aerospace field are mostly for specific equipment, and scholars mostly assess the real-time health status of launch vehicle engines which can only reflect the current health status of the launch vehicle. Existing methods cannot be applied to different equipment, and there is a lack of research on health assessments of fuzzy and complex mechanical systems. In this article, we propose a multi-layer and multi-factor predictive evaluation method for a fuzzy and complex system and conduct experiments on real vibration data of rockets. First, we divide the health assessment level according to the vibration data that affect the normal operation of the rocket. Secondly, we obtain the future trend of vibration signals based on five data prediction methods and calculate the health status interval of the rocket engine's working conditions based on the boxplot method. At the same time, we calculate the single health evaluation set of every vibration signal. We obtain the weights of each level and factor for the health value based on an analytic hierarchy process (AHP). The optimization of this step avoids an over-reliance on expert experience. Finally, we complete a fuzzy comprehensive evaluation of the engine system from the bottom up to obtain the final health value. The minimum evaluation error is 0.0193% on the test data of the Long March series launch vehicle engine, which shows that the proposed method can successfully predict and evaluate the launch vehicle engine. [ABSTRACT FROM AUTHOR]

The spreading of disinformation in social media threatens cybersecurity and undermines market efficiency. Detecting disinformation is challenging due to large volumes of social media content and a rapidly changing environment. This research developed and validated a theory-based, novel deep-learning approach (called TRNN) to disinformation detection. Grounded in social and psychological theories, TRNN uses deep-learning and data-centric augmentation to enhance disinformation detection in financial social media. Temporal and contextual information is encoded as specific knowledge about human-validated disinformation, which was identified from our unique collection of 745,139 financial social media messages about four U.S. high-tech company stocks and their fine-grained trading data. TRNN uses multiple series of long short-term memory (LSTM) recurrent neurons to learn dynamic and hidden patterns to support disinformation detection. Our experimental findings show that TRNN significantly outperformed widely-used machine learning techniques in terms of precision, recall, F-score and accuracy, achieving consistently better classification performance in disinformation detection. A case study of Apple Inc.'s stock price movement demonstrates the potential usability of TRNN for secure knowledge management. The research contributes to developing novel approach and model, producing new information systems artifacts and dataset, and providing empirical findings of detecting online disinformation. [ABSTRACT FROM AUTHOR]

Antifreeze protein (AFP) can inhibit the growth of ice crystals to protect organisms from freezing damage, and demonstrates broad application prospects in food industry. Antifreeze peptides (AFPP) are specific peptides with functional domains showing antifreeze activity in AFP. Bioinformatics-based molecular simulation technology can more accurately explain the properties and mechanisms of biological macromolecules. Therefore, the binding stability of antifreeze peptides and antifreeze proteins (AFP(P)) to ice and the molecular-scale growth kinetics of ice were analyzed by molecular simulation, which can make up for the limitations of experimental technology. This review concludes the molecular simulation-based research in the inhibition's study of AFP(P) on ice growth, including sequence prediction, structure construction, molecular docking and molecular dynamics (MD) studies of AFP(P) on ice applications in growth inhibition. Finally, the review prospects the future direction of designing new antifreeze biomimetic materials through molecular simulation and machine learning. The information presented in this paper will help enrich our understanding of AFPP.

Abstract Alzheimer’s disease (AD) is a slowly progressing disease for which there is no known therapeutic cure at present. Ongoing research around the world is actively engaged in the quest for identifying markers that can help predict the future cognitive state of individuals so that measures can be taken to prevent the onset or arrest the progression of the disease. Researchers are interested in both biological and neuropsychological markers that can serve as good predictors of the future cognitive state of individuals. The goal of this study is to identify non-invasive, inexpensive markers and develop neural network models that learn the relationship between those markers and the future cognitive state. To that end, we use the renowned Alzheimer’s Disease Neuroimaging Initiative (ADNI) data for a handful of neuropsychological tests to train Recurrent Neural Network (RNN) models to predict future neuropsychological test results and Multi-Level Perceptron (MLP) models to diagnose the future cognitive states of trial participants based on those predicted results. The results demonstrate that the predicted cognitive states match the actual cognitive states of ADNI test subjects with a high level of accuracy. Therefore, this novel two-step technique can serve as an effective tool for the prediction of Alzheimer’s disease progression. The reliance of the results on inexpensive, non-invasive tests implies that this technique can be used in countries around the world including those with limited financial resources.

Smart homes have been recently important sources for providing Activity of Daily Living (ADL) data about their residents. ADL data can be a great asset while analyzing residents’ behavior to provide residents with better and optimized services. A popular example is to analyze residents’ behavior to predict their future activities and optimize smart homes performance accordingly. This paper proposes a forecasting framework that utilizes ADL data to predict residents’ next activities in a smart home environment. Forecasting is performed via the conjunction of embedding algorithm to encode the data and Bidirectional Long Short-Term Memory (BiLSTM) deep neural networks to process the data. The proposed framework is evaluated over five real ADL datasets where the experiments show the outperformance of the proposed framework with accuracy scores ranging from 98.7% to 93.8%.

In this paper, we propose a unified deep learning model for monitoring elderly in execution of daily life activities such as eating, sleeping or taking medication. The proposed approach consists of three stages which are activity recognition, anomaly detection and next activity prediction. Such a system can provide useful information for the elderly, caregivers and medical teams to identify activities and generate preventive and corrective measures. In literature, these stages are discussed separately, however, in our approach, we make use of each stage to progress into the next stage. At first, activity recognition based on different extracted features is performed using a deep neural network (DNN), then an overcomplete-deep autoencoder (OCD-AE) is employed to separate the normal from anomalous activities. Finally, a cleaned sequence of consecutive activities is constructed and used by a long short-term memory (LSTM) algorithm to predict the next activity. Since the last two stages depend on the activity recognition stage, we propose to increase its accuracy by exploiting different extracted features. The performance of the proposed unified approach has been evaluated on real smart home datasets to demonstrate its ability to recognize activities, detect anomalies and predict the next activity.

Precipitation nowcasting is an important but hard problem. Currently, with the landing of deep learning, it has been treated as an image prediction problem based on radar echo maps. However, deep learning models suffer from poor performance and blurred prediction results. Lots of improvement works enhance the model by adding complex modules, which increases insufferable training memory and time overhead. Others tempt to add more limitations or guidances on loss, but they usually have little effect in such an extremely complex and difficult task. In this article, we propose a simple but effective loss named focal frame loss (FFL), which assigns different weights to the images in the prediction sequence to focus on the images that are relatively difficult to predict. Experiments on two large-scale radar datasets show that FFL can greatly improve the performance of multiple popular models without introducing additional training costs.

The prediction of express delivery sequence, i.e., modeling and estimating the volumes of daily incoming and outgoing parcels for delivery, is critical for online business, logistics, and positive customer experience, and specifically for resource allocation optimization and promotional activity arrangement. A precise estimate of consumer delivery requests has to involve sequential factors such as shopping behaviors, weather conditions, events, business campaigns, and their couplings. Despite that various methods have integrated external features to enhance the effects, extant works fail to address complex feature-sequence couplings in the following aspects: weaken the inter-dependencies when processing heterogeneous data and ignore the cumulative and evolving situation of coupling relationships. To address these issues, we propose DeepExpress—a deep-learning-based express delivery sequence prediction model, which extends the classic seq2seq framework to learn feature-sequence couplings. DeepExpress leverages an express delivery seq2seq learning, a carefully designed heterogeneous feature representation, and a novel joint training attention mechanism to adaptively handle heterogeneity issues and capture feature-sequence couplings for accurate prediction. Experimental results on real-world data demonstrate that the proposed method outperforms both shallow and deep baseline models. [ABSTRACT FROM AUTHOR]

Electrical load forecasting is one of the critical tasks that helps power utility companies in planning and operation as well as the energy managementsystem (EMS) in controlling and optimizing the power grid’s performance. It guarantees system adequacy and reliability by reducing the gap between supply and demand, and it bolsters cybersecurity by enabling false data detection and system security by predicting system peak load. This article aims to build an accurate long-term load forecasting model based on deep learning (DL) algorithms. Initially, big-data analytics and feature engineering are applied to the power load and weather historical data. Next, three different load forecasting models are developed based on three corresponding deep neural network models (DNN): the simple RNN, the long short term memory (LSTM), and the gated recurrent unit (GRU), respectively. To find an accurate DNN model, a two-step hyperparameter optimization is performed consisting of cross-validation. In the first step, the random search (RS) algorithm is employed to narrow the parameter search space to a smaller local one. In the second step, the grid search (GS) algorithm is used to find the best hyperparameters in the local area located by RS. However, the computational burden of the DNN training process is too high and may lead to out-of-memory (OOM) issues, especially when we are dealing with a big volume of data. To overcome this problem, we developed our Python program based on Tensorflow framework and implemented the framework on a high-performance computer cluster (HPCC). Then, different jobs are defined to employ parallel distributed computing (PDC) capability on the SIU BigDawg HPCC. To this end, all DNN hyperparameters (e.g., number of layers, neurons, dropout, batch size, epoch, learning rate, etc.) can be optimized and result in a comprehensive and accurate load forecasting model. Our proposed method, along with the programming framework, is evaluated on a real-world dataset. Among the three models developed and tested, the GRU, with load and weather as features, has demonstrated excellent performance and outperformed the other two models for long-term forecasting with the best metrics (i.e., MSE, MAPE, $R^2$) and daily error almost less than 1%. Also, our best model (optimized GRU) is compared with the other reported state-of-the-art models such as DBN, SVR, RM-LSTM, DM-GCNN, TCMS-CNN. It is demonstrated that the proposed framework yields a superior performance than the others. [ABSTRACT FROM AUTHOR]

Accurate and timely nowcasting precipitation has huge social and economic benefits. However, the changes in clouds including expansion, dissipation, and distortion are extremely complex, which exacerbates the difficulty of forecasting. Fortunately, it still follows certain meteorological laws, which can be explored based on spatiotemporal information but are not fully considered by previous models. In this article, we design two modules to focus on these messages based on atmospheric characteristics. Specifically, the spatial local attention memory (SLAM) module combines local attention and memory mechanism to capture the meteorological spatial relationship, while the time difference memory (TDM) module combines differential technology and memory mechanism to capture the meteorological temporal variants. We combine these two modules with PredRNN and propose PrecipLSTM to sufficiently capture the spatiotemporal dependencies of radar data. We do exhaustive experiments with five baselines on four radar datasets. It is verified that PrecipLSTM achieves state-of-the-art results with fewer parameters than the previous state-of-the-art method. [ABSTRACT FROM AUTHOR]

G-quadruplexes (G4s) have been long considered rare and physiologically unimportant in vitro curiosities, but recent methodological advances have proved their presence and functions in vivo. Moreover, in addition to their functional relevance in bacteria and animals, including humans, their importance has been recently demonstrated in evolutionarily distinct plant species. In this study, we analyzed the genome of Pisum sativum (garden pea, or the so-called green pea), a unique member of the Fabaceae family. Our results showed that this genome contained putative G4 sequences (PQSs). Interestingly, these PQSs were located nonrandomly in the nuclear genome. We also found PQSs in mitochondrial (mt) and chloroplast (cp) DNA, and we experimentally confirmed G4 formation for sequences found in these two organelles. The frequency of PQSs for nuclear DNA was 0.42 PQSs per thousand base pairs (kbp), in the same range as for cpDNA (0.53/kbp), but significantly lower than what was found for mitochondrial DNA (1.58/kbp). In the nuclear genome, PQSs were mainly associated with regulatory regions, including 5′UTRs, and upstream of the rRNA region. In contrast to genomic DNA, PQSs were located around RNA genes in cpDNA and mtDNA. Interestingly, PQSs were also associated with specific transposable elements such as TIR and LTR and around them, pointing to their role in their spreading in nuclear DNA. The nonrandom localization of PQSs uncovered their evolutionary and functional significance in the Pisum sativum genome. [ABSTRACT FROM AUTHOR]

Sixty percent of the failures of launch vehicles in the ascending phase occur in the propulsion system. Among them, the vibration generated by the engine is an important factor in the occurrence of failure. At present, health assessment methods in the aerospace field are mostly for specific equipment, and scholars mostly assess the real-time health status of launch vehicle engines which can only reflect the current health status of the launch vehicle. Existing methods cannot be applied to different equipment, and there is a lack of research on health assessments of fuzzy and complex mechanical systems. In this article, we propose a multi-layer and multi-factor predictive evaluation method for a fuzzy and complex system and conduct experiments on real vibration data of rockets. First, we divide the health assessment level according to the vibration data that affect the normal operation of the rocket. Secondly, we obtain the future trend of vibration signals based on five data prediction methods and calculate the health status interval of the rocket engine’s working conditions based on the boxplot method. At the same time, we calculate the single health evaluation set of every vibration signal. We obtain the weights of each level and factor for the health value based on an analytic hierarchy process (AHP). The optimization of this step avoids an over-reliance on expert experience. Finally, we complete a fuzzy comprehensive evaluation of the engine system from the bottom up to obtain the final health value. The minimum evaluation error is 0.0193% on the test data of the Long March series launch vehicle engine, which shows that the proposed method can successfully predict and evaluate the launch vehicle engine.

Opinion dynamics (OD) models simulate the evolution of personal opinions using computer science, social dynamics, physics, and complexity science tools on social networks. However, current opinion dynamics models mainly utilize agent-based models to simulate user interactions but still need to leverage large-scale data fully. Social networks typically involve a vast number of users and information, making it difficult for models to capture the complex interactions in real social networks. Additionally, many models overlook the specific mechanisms of user interactions, such as how network individuals influence and interact with one another, resulting in inaccuracies in modeling individual opinion evolution. This paper proposes a novel approach that integrates the probabilistic bounded confidence model (PBC) with deep learning, i.e., Neural Probabilistic Bounded Confidence (NPBC), to tackle these challenges. The NPBC model adopts an attention network, a long short-term memory (LSTM) network, and a feedforward neural network to approximate the evolution of individual opinions within networks. It satisfies the data approximation and the users' opinion update rule that conforms to the PBC model. This model enhances the understanding of opinion dynamics, helping to develop effective strategies for opinion guidance in social networks. Finally, the proposed NPBC model is evaluated on three synthetic and Twitter datasets. The results demonstrate that the proposed NPBC model performs better than the six baseline methods in predicting user opinions. Compared to baseline models, the model's accuracy (ACC) and comprehensive evaluation (F1) scores improve by up to 18% and 7%, respectively. • Combining deep learning and probabilistic bounded confidence model. • Using fully connected layers to map users and time to the same vector space. • The attention mechanism captures the internal correlations between users and time. • The long short-term memory network is used to capture the sequence features. [ABSTRACT FROM AUTHOR]

In structural health monitoring (SHM), measuring and evaluating structural dynamic responses are critical for safety management of civil infrastructures. Particularly, online forecasting of the structural responses under extreme external loading conditions (e.g., earthquakes) takes a significant role in SHM to provide early warning and ensure safe operation. In practice, complex causality and intrinsic interactions between seismic excitation and structural response make it challenging to establish a reliable predictive scheme. The present paper proposes a novel deep recurrent neural network (RNN) model implemented in the architecture of a time-series attention-based RNN encoder–decoder (TSA-RNN-ED), for predictive analysis of structural responses under seismic excitation. In the proposed data-driven model, upcoming sequential responses are predicted through sequence-to-sequence learning from historical multivariate time-series signals. A time-series attention mechanism is proposed to exploit the heterogeneous, but directly related, hidden features between the seismic loads and the corresponding structural responses. The proposed architecture can reliably regress excitation-response interactions to predict dynamic responses subjected to future earthquakes while satisfying the need of real-time forecasting for on-site practical implementation. This article systematically evaluates the proposed model by using two real-world structural cases: 1) the tallest building in China, the Shanghai Tower and 2) a woodframe classroom on a shake table at the University of British Columbia in Vancouver, Canada. The experimental results demonstrate the accurate and efficient performance of the proposed methodology in forecasting the seismic responses of the structures under investigation. [ABSTRACT FROM AUTHOR]

Alzheimer’s disease (AD) is a slowly progressing disease for which there is no known therapeutic cure at present. Ongoing research around the world is actively engaged in the quest for identifying markers that can help predict the future cognitive state of individuals so that measures can be taken to prevent the onset or arrest the progression of the disease. Researchers are interested in both biological and neuropsychological markers that can serve as good predictors of the future cognitive state of individuals. The goal of this study is to identify non-invasive, inexpensive markers and develop neural network models that learn the relationship between those markers and the future cognitive state. To that end, we use the renowned Alzheimer’s Disease Neuroimaging Initiative (ADNI) data for a handful of neuropsychological tests to train Recurrent Neural Network (RNN) models to predict future neuropsychological test results and Multi-Level Perceptron (MLP) models to diagnose the future cognitive states of trial participants based on those predicted results. The results demonstrate that the predicted cognitive states match the actual cognitive states of ADNI test subjects with a high level of accuracy. Therefore, this novel two-step technique can serve as an effective tool for the prediction of Alzheimer’s disease progression. The reliance of the results on inexpensive, non-invasive tests implies that this technique can be used in countries around the world including those with limited financial resources. [ABSTRACT FROM AUTHOR]

The article analyses the research related to the issue of detecting anomalies in current medical data. The work aims to develop an information system for detecting data anomalies in the format of time series, such as metrics, with the ability to visualize the results for expert evaluation. The process of detecting anomalies by machine learning tools for detecting anomalies in the flow data is investigated. A system for detecting anomalies in metrics using the HTM model is built. According to the model results, obtained satisfactory accuracy for detecting anomalies at standard network parameters, the estimation of anomalies differed clearly for different aggregation intervals. The OPF Client functionality was used to build the HTM model, which allowed to achieve speed and simplicity in its construction. The obtained results allow the expert to choose the best of the size of the intervals as well as the parameters of the models. This choice is necessary because the presence of the anomaly depends on expert knowledge in a particular field. [ABSTRACT FROM AUTHOR]

This article presents a novel methodology to detect insurance claim related frauds in the healthcare system using concepts of sequence mining and sequence prediction. Fraud detection in healthcare is a non-trivial task due to the heterogeneous nature of healthcare records. Fraudsters behave as normal patients and with the passage of time keep on changing their way of planting frauds; hence, there is a need to develop fraud detection models. The sequence generation is not the part of previous researches which mostly focus on amount based analysis or medication versus diseases sequential analysis. The proposed methodology is able to generate sequences of services availed or prescribed by each specialty and analyse via two cascaded checks for the detection of insurance claim related frauds. The methodology addresses these challenges and self learns from historical medical records. It is based on two modules namely “Sequence rule engine and Prediction based engine”. The sequence rule engine generates frequent sequences and probabilities of rare sequences for each specialty of the hospital. The comparison of such sequences with the actual patient sequences leads to the identification of anomalies as both sequences are not compliant to the sequences of the rule engine. The system performs further in detail analysis on all non-compliant sequences in the prediction based engine. The proposed methodology is validated by generating patient sequences from last five years transactional data of a local hospital and identifies patterns of service procedures administered to patients using Prefixspan algorithm and Compact prediction tree. Various experiments have been performed to validate the applicability of the developed methodology and the results demonstrate that the methodology is pertinent to detect healthcare frauds and provides on average 85% of accuracy. Thus can help in preventing fraudulent claims and provides better insight into how to improve patient management and treatment procedures.

This paper proposes a region-based travel time and traffic speed prediction method using sequence prediction. Floating Car Data collected from 8,317 vehicles during 34 days are used for evaluation purposes. Twelve districts are chosen and the spatio-temporal non-linear relations are learned with Recurrent Neural Networks. Time estimation of the total trip is solved by travel time estimation of the divided sub-trips, which are constituted between two consecutive GNSS measurement data. The travel time and final speed of sub-trips are learned with Long Short-term Memory cells using sequence prediction. A sequence is defined by including the day of the week meta-information, dynamic information about vehicle route start and end positions, and average travel speed of the road segment that has been traversed by the vehicle. The final travel time is estimated for this sequence. The sequence-based prediction shows promising results, outperforms function mapping and non-parametric linear velocity change based methods in terms of root-mean-square error and mean absolute error metrics.

Laboratory performance testing of the phase angle of asphalt concrete (AC) mixtures is often expensive and requires enormous human effort and time. To circumvent this problem, several regression-based methods have been proposed in the literature to model the phase angle behavior of AC mixtures using various approaches. However, these methods impose strict assumptions on the underlying relationship between phase angle and its corresponding covariates as well as how well and accurately these covariates are measured, restricting us from fully analyzing the predictive capability of any modeling method. To this end, this study proposed an alternative approach for modeling the phase angle characteristics of AC mixtures based on a recurrent neural network (RNN) that inherently and implicitly captures the effects of covariates. This approach is suitable to model the sequential nature of data recorded in laboratory testing where phase angle testing was repeated for a set of six loading frequencies forming a recurrent pattern. The proposed RNN model (P-RNN) was applied separately to wearing and base course mixtures by considering the historical values of phase angle as input and to predict its value for the next loading frequency, keeping temperature as a constant. To demonstrate the superiority of the proposed approach, the P-RNN model is compared with other competing models from the literature, and the results reveal superior performance of the P-RNN model. [ABSTRACT FROM AUTHOR]

This article is about a challenging problem called matrix sequence prediction, which is motivated from the application of taxi order prediction. Remarkably, the problem differs greatly from previous sequence prediction tasks in the sense that the time-wise correlations are quite elusive; namely, distant entries could be strongly correlated and nearby entries are unnecessarily related. Such distinct specifics make prevalent convolution-recurrence-based methods inadequate to apply. To remedy this trouble, we propose a novel architecture called Parallel Connected LSTM (PcLSTM), which integrates two new mechanisms, Multi-channel Linearized Connection (McLC) and Adaptive Parallel Unit (APU), into the framework of LSTM. Benefiting from the strengths of McLC and APU, our PcLSTM is able to handle well both the elusive correlations within each timestamp and the temporal dependencies across different timestamps, achieving state-of-the-art performance in a set of experiments demonstrated on synthetic and real-world datasets. [ABSTRACT FROM AUTHOR]

With the rapid development of artificial intelligence and industrial Internet of Things (IIoT) technologies, intelligent predictive maintenance (IPdM) has received considerable attention from researchers and practitioners. To efficiently predict impending failures and mitigate unexpected downtime, while satisfying the instant maintenance demands of industrial facilities is very important for improving the production efficiency. In this article, a self-attention based “Perception and Prediction” framework, called DeepHealth, is proposed for the instant IPdM. Specifically, the framework is composed of two submodels (i.e., DH-1 and DH-2), which are respectively utilized to perform the health perception and sequence prediction. By operating the framework, the proposed models can predict the health conditions via predicting the future signal samples, thereby completing the instant IPdM. Considering the potential temporal correlation in time series, we deploy an enhanced attention mechanism to capture global dependencies from the vibration signals, and leverage the long- and short-term sequence prediction of sensor signals to support instant maintenance decision-making. On this basis, we conduct a destructive experiment based on the IIoT-enabled rotating machinery and construct a balanced industrial dataset for model evaluations. Extensive experiment results show that the proposed solution achieves good prediction accuracy for instant IPdM on the automatic washing equipment and Case Western Reserve University datasets. [ABSTRACT FROM AUTHOR]

An automated smart home system is a key contributor to user assistance technology in modern civilization. Crucial merit of such a system is its ability to train itself through recorded data and recognize patterns in resident behaviors. Lack of sufficient prediction accuracy, exponential memory consumption, and extensive runtime prevent many of the current activity prediction approaches from being seamlessly integrated into consumer residences. This research introduces a sequence prediction algorithm which uses a prefix tree-based data model in order to learn and predict user actions. The algorithm applies episode discovery to detect correlated sensor events and learns the activities using a lossless data compression technique. This process assigns a probability of occurrence to sensor events and uses these probabilities to detect patterns in resident behavior. A complexity analysis of the algorithm is done to prove its efficiency in terms of memory usage and runtime. Using the presented technique, predictions are performed on popular datasets and contrasted with existing algorithms. The proposed algorithm achieves an 8.22% improvement in prediction accuracy over its predecessors, along with 66.69% better memory efficiency and 37% faster runtime.

Discovering unseen patterns from web clickstream is an upcoming research area. One of the meaningful approaches for making predictions is using sequence prediction that is typically the improved compact prediction tree (CPT+). However, to increase this method's effectiveness, combining it with at least other methods is necessary. This work investigates such PageRankbased methods related to sequence prediction as All-K-Markov, DG, Markov 1st, CPT, CPT+. The experimental results proved that the integration of CPT+ and PageRank is the right solution for next page prediction in terms of accuracy, which is more than a standard method of approximately 0.0621%. Still, the size of the newly created sequence database is reduced up to 35%. Furthermore, our proposed solution has an accuracy that is much higher than other ones. It is intriguing for the next phase (testing one) to make the next page prediction in terms of time performance. [ABSTRACT FROM AUTHOR]

Information search in video streaming services such as YouTube is replacing traditional information search services. To find desired detailed information in such a video, users should repeatedly navigate several points in the video, resulting in a waste of time and network traffic. In this paper, we propose a method to assist users in searching for information in a video by using DBSCAN clustering and LSTM. Our LSTM model is trained with a dataset that consists of user search sequences and their final target points categorized by DBSCAN clustering algorithm. Then, our proposed method utilizes the trained model to suggest an expected category for the user's desired target point based on a partial search sequence that can be collected at the beginning of the search. Our experiment results show that the proposed method successfully finds user destination points with 98% accuracy and 7s of the time difference by average. [ABSTRACT FROM AUTHOR]

Abstract Background MicroRNAs (miRNAs) are small, non-coding RNA molecules that play a key role in gene regulation in both plants and animals. MicroRNA biogenesis involves the enzymatic processing of a primary RNA transcript. The final step is the production of a duplex molecule, often designated as miRNA:miRNA*, that will yield a functional miRNA by separation of the two strands. This miRNA will be incorporated into the RNA-induced silencing complex, which subsequently will bind to its target mRNA in order to suppress its expression. The analysis of miRNAs is still a developing area for computational biology with many open questions regarding the structure and function of this important class of molecules. Here, we present StarSeeker, a simple tool that outputs the putative miRNA* sequence given the precursor and the mature sequences. Results We evaluated StarSeeker using a dataset consisting of all plant sequences available in miRBase (6992 precursor sequences and 8496 mature sequences). The program returned a total of 15,468 predicted miRNA* sequences. Of these, 2650 sequences were matched to annotated miRNAs (~ 90% of the miRBase-annotated sequences). The remaining predictions could not be verified, mainly because they do not comply with the rule requiring the two overhanging nucleotides in the duplex molecule. Conclusions The expression pattern of some miRNAs in plants can be altered under various abiotic stress conditions. Potential miRNA* molecules that do not degrade can thus be detected and also discovered in high-throughput sequencing data, helping us to understand their role in gene regulation.

The prediction of next location for users in location-based social networks has become an increasing significant requirement since it can benefit both users and business. However, existing methods lack an integrated analysis of sequence context, input contexts, and user preferences in a unified way, and result in an unsatisfactory prediction. Moreover, the interaction between different kinds of input contexts has not been investigated. In this paper, we propose a multi-context integrated deep neural network model (MCI-DNN) to improve the accuracy of the next location prediction. In this model, we integrate sequence context, input contexts, and user preferences into a cohesive framework. First, we model sequence context and interaction of different kinds of input contexts jointly by extending the recurrent neural network to capture the semantic pattern of user behaviors from check-in records. After that, we design a feedforward neural network to capture high-level user preferences from check-in data and incorporate that into MCI-DNN. To deal with different kinds of input contexts in the form of multi-field categorical, we adopt embedding representation technology to automatically learn dense feature representations of input contexts. Experimental results on two typical real-world data sets show that the proposed model outperforms the current state-of-the-art approaches by about 57.12% for Foursquare and 76.4% for Gowalla on average regarding F1-score@5.

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Data of sequential nature arise in many application domains in the form of, e.g., textual data, DNA sequences, and software execution traces. Different research disciplines have developed methods to learn sequence models from such datasets: (i) In the machine learning field methods such as (hidden) Markov models and recurrent neural networks have been developed and successfully applied to a wide range of tasks, (ii) in process mining process discovery methods aim to generate human-interpretable descriptive models, and (iii) in the grammar inference field the focus is on finding descriptive models in the form of formal grammars. Despite their different focuses, these fields share a common goal: learning a model that accurately captures the sequential behavior in the underlying data. Those sequence models are generative, i.e., they are able to predict what elements are likely to occur after a given incomplete sequence. So far, these fields have developed mainly in isolation from each other and no comparison exists. This paper presents an interdisciplinary experimental evaluation that compares sequence modeling methods on the task of next-element prediction on four real-life sequence datasets. The results indicate that machine learning methods, which generally do not aim at model interpretability, tend to outperform methods from the process mining and grammar inference fields in terms of accuracy. [ABSTRACT FROM AUTHOR]

• A hierarchical temporal attention based model is proposed to support short-term and long-term human mobility sequence prediction. • The proposed hierarchical temporal attention incorporates individual mobility patterns into the model architecture. • The model is compared with four baseline methods on individual trajectory datasets with varying degree of traveling uncertainty. • Experiments demonstrate the outperformance of the proposed method using three evaluation metrics. • The proposed model uncovers individual frequential and periodical mobility patterns in an interpretable manner. Prediction of individual mobility is crucial in human mobility related applications. Whereas, existing research on individual mobility prediction mainly focuses on next location prediction and short-term dependencies between traveling locations. Long-term location sequence prediction is of great importance for long-time traffic planning and location advertising, and long-term dependencies exist as individual mobility regularity typically occurs daily and weekly. This paper proposes a novel hierarchical temporal attention-based LSTM encoder-decoder model for individual location sequence prediction. The proposed hierarchical attention mechanism captures both long-term and short-term dependencies underlying in individual longitudinal trajectories, and uncovers frequential and periodical mobility patterns in an interpretable manner by incorporating the calendar cycle of individual travel regularities into location prediction. More specifically, the hierarchical attention consists of local temporal attention to identify highly related locations in each day, and global temporal attention to discern important travel regularities over a week. Experiments on individual trajectory datasets with varying degree of traveling uncertainty demonstrate that our method outperforms four baseline methods on three evaluation metrics. In addition, we explore the interpretability of the proposed model in understanding individual daily, and weekly mobility patterns by visualizing the temporal attention weights and frequent traveling patterns associated with locations. [ABSTRACT FROM AUTHOR]

The inevitable success of predictive analytics lies in the prospect of cutting edge approaches proposed in the domain of data mining and machine learning. Frequent itemsets, sequential pattern mining has been the major sub domains of the data mining since its inception with human work. Sequence prediction is a rather more targeted approach towards the predictive data mining. Sequential pattern mining approaches, while being very efficient in finding associations and interactions of difference itemsets and discovering hidden pattern in the transactional or graph datasets, sequence prediction is better focused on predicting the next item of an existent sequence. Applications of sequence prediction has been in pattern prediction, stock value prediction, websites prefetching, fraud detection, breach prediction, text prediction, product recommendation in extensive manner. Several state-of-the-art algorithms and models have been proposed starting from graphical models, markovian models and tree based structures have been proposed by data mining. Machine learning models like recurrent neural networks with LSTM blocks work pretty well in terms of performance. Our aim is to propose an integrated data structure with better performance measures (space-time efficiency, scalability) and which can be effectively used in online learning setting-stream processing algorithm. The paper also describes the extension prospect of this algorithm which can be used in the time-series forecast. [ABSTRACT FROM AUTHOR]

Distribution feeder long-term load forecast (LTLF) is a critical task many electric utility companies perform on an annual basis. The goal of this task is to forecast the annual load of distribution feeders. The previous top-down and bottom-up LTLF methods are unable to incorporate different levels of information. This paper proposes a hybrid modeling method using sequence prediction for this classic and important task. The proposed method can seamlessly integrate top-down, bottom-up, and sequential information hidden in multi-year data. Two advanced sequence prediction models long short-term memory (LSTM) and gated recurrent unit (GRU) networks are investigated in this paper. They successfully solve the vanishing and exploding gradient problems a standard recurrent neural network has. This paper first explains the theories of LSTM and GRU networks and then discusses the steps of feature selection, feature engineering and model implementation in detail. In the end, a real-world application example for a large urban grid in West Canada is provided. LSTM and GRU networks under different sequential configurations and traditional models including bottom-up, ARIMA, and feed-forward neural network are all implemented and compared in detail. The proposed method demonstrates superior performance and great practicality. [ABSTRACT FROM AUTHOR]

Similarity-based approaches have made significant progress in visual object tracking (VOT). Although these methods work well in simple scenes, they ignore the continuous spatio-temporal connection of the object in the video sequence. For this reason, tracking by spatial matching solely can lead to tracking failures because of distractors and occlusion. In this paper, we propose a spatio-temporal joint-modeling tracker named STTrack which implicitly builds continuous connections between the temporal and spatial aspects of the sequence. Specifically, we first design a time-sequence iteration strategy (TSIS) to concentrate on the temporal connection of the object in the video sequence. Then, we propose a novel spatial temporal interaction Transformer network (STIN) to capture the spatio-temporal correlation of the object between frames. The proposed STIN module is robust in object occlusion because it explores the dynamic state change dependencies of the object. Finally, we introduce a spatio-temporal query to suppress distractors by iteratively propagating the target prior. Extensive experiments on six tracking benchmark datasets demonstrate that the proposed STTrack achieves excellent performance while operating in real-time. The code is publicly available at https://github.com/nubsym/STTrack. [ABSTRACT FROM AUTHOR]

The problem is that of sequential probability forecasting for finite-valued time series. The data is generated by an unknown probability distribution over the space of all one-way infinite sequences. Two settings are considered: the realizable and the non-realizable one. Assume first that the probability measure generating the sequence belongs to a given set C (realizable case), but the latter is completely arbitrary (uncountably infinite, without any structure given). It is shown that the minimax asymptotic average loss which may be positive is always attainable, and it is attained by a Bayesian predictor whose prior is discrete and concentrated on C. Moreover, the finite-time loss of the Bayesian predictor is also optimal up to an additive log n term (where n is the time step). This upper bound is complemented by a lower bound that goes to infinity but may do so arbitrarily slow. Passing to the non-realizable setting, let the probability measure generating the data be arbitrary, and consider the given set C as a set of experts to compete with. The goal is to minimize the regret with respect to the experts. It is shown that in this setting it is possible that all Bayesian strategies are strictly suboptimal even asymptotically. In other words, a sublinear regret may be attainable but the regret of every Bayesian predictor is linear. A very general recommendation for choosing a model can be made based on these results: it is better to take a model large enough to make sure it includes the process that generates the data, even if it entails positive asymptotic average loss, for otherwise any combination of predictors in the model class may be useless. [ABSTRACT FROM AUTHOR]

Natural language processing is a technique to process data such as text and speech. Some fundamental research includes named-entity recognition, which recognizes name entities (i.e., persons, companies) from texts; semantic parsing, which is used to convert a natural language utterance to the representation of logical form; and co-reference resolution, which extracts nouns (including pronouns, noun phrases) pointing to the same reference body. In this paper, we mainly focus on the task of mention extraction, which extract and classify overlapping or nested structure mentions. We proposed a neural-encoded mention-hypergraph (NEMH) model to use hypergraph to model overlapping or nested structure mentions and use neural networks to extract features for hypergraph automatically. Unlike the existing approaches, our hypergraph model can effectively capture nested mention entities with unlimited lengths. Also, the proposed model is highly scalable and the time complexity of the proposed model is linear in the number of mention classes and the number of input words. Extensive experiments are conducted on several standard datasets to demonstrate the effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]

Highlights • A three step combination of different Neural Networks for discrimination. • The algorithm works by decomposing and remembering previously presented images. • The training of each of the networks is separated according to each of its properties. • The algorithm perform better at discrimination than individual methods. • The algorithm can remember longer sequences than individual methods. Abstract Although done nearly effortlessly by humans, digital systems cannot easily recognize images or predictions from recent observations. Tackling these limitations by proposing novel algorithms to improve the performance of image processing would have widespread implications in a variety of fields, including robotics, manufacturing, biomedicine, and automation. To provide a computer with this combined ability and transform it into an intelligent system, an algorithm must combine memory with an image decomposition procedure. Artificial neural networks (ANNs) are algorithms that aim to solve tasks such as classification, clustering, pattern recognition, and prediction by resembling brain connections. Specifically, three ANNs have excelled in specific areas: deep neural networks (DNNs), which use intrinsic connections to create prediction maps; long short-term memory neural networks (LSTMs), which use recurrent connections to emulate a type of memory; and convolutional neural networks (CNNs), which can decompose complex data through layers for simpler analysis. Although these algorithms can solve certain tasks of image sequence prediction, they cannot easily solve entire problems on their own. Nevertheless, combining these networks may enable solving such problems with ease. Thus, this article evaluates the combination of ANNs into two novel algorithms developed with the aim of improving image sequence prediction: (i) a combination of CNNs and LSTMs to form a CLNN and (ii) a combination of CNNs, LSTMs, and DNNs to form a CLDNN. Although the developed algorithms require a longer training time, they require less training epochs to have better accuracy than their predecessors. Furthermore, both developed methods were capable of accurately performing the image sequence prediction task, outperforming each individual method, as well as predicting longer and greater numbers of sequences correctly. Overall, the developed algorithms were able to better decompose inputs, remember previous inputs, and more accurately predict sequences of images. This allows the prediction of the next step in the sequence, which can be used as part of an intelligent system to make an analysis and an informed decision on the next course of action. [ABSTRACT FROM AUTHOR]