Modeling Sequential Data Research Articles

Motor Bearing Failure Identification Using Multiple Long Short-Term Memory Training Strategies

In the context of condition-based maintenance of rotating machines in manufacturing systems, the early diagnosis of possible faults related to rolling elements of the bearing is mainly based on techniques from artificial intelligence, namely, Machine Learning (ML) and Deep Learning (DL). Approaches based on using Deep Learning methods have been the most coveted in recent years. Among a variety of models, the type of architecture known as Long-Short-Term Memory (LSTM) of Recurrent Neural Network (RNN) has both the ability to capture long-term dependencies and to adapt to sequential data modeling. It is therefore able to work on data without any preprocessing. This paper studies using four types of LSTM networks to diagnose bearing faults in a classification approach. It aims to intervene on both the input parameters and the network architecture, to achieve high performance. The proposed method is carried out in two different ways. In the first case, the data inputs are raw frames of vibration signals. However, in the second case, the network inputs are pre-computed time-frequency features. The results clearly showed that LSTMs are more accurate with the latter.

Prediction of Solar Radiation using Deep LSTM-based Machine Learning Algorithm

Solar radiation, a critical parameter for various applications such as solar energy systems and weather forecasting, exhibits complex temporal patterns influenced by numerous environmental factors. In the quest to enhance the accuracy of solar radiation predictions, this study introduces a novel approach utilizing Deep Long Short-Term Memory (Deep LSTM) networks, a type of recurrent neural network (RNN) known for its capability to model sequential data. Traditional prediction methods often struggle to capture these intricacies, leading to sub-optimal performance. This investigation is based on a Deep LSTM-based machine learning algorithm designed to predict solar radiation with elevated exactitude. The model is meticulously trained on copious datasets encompassing historical meteorological data, including temperature, humidity, wind velocity, and antecedent solar radiation metrics. Pivotal stages encompass data preprocessing, judicious feature selection, and hyperparameter optimization to enhance the model’s predictive efficacy. Empirical results clearly illustrate that the Deep LSTM model surpasses traditional methodologies, attaining superior accuracy and resilience across diverse meteorological scenarios. The ramifications of this endeavor portend significant advancements in the strategic planning and administration of solar energy resources, thereby contributing to more dependable and efficacious renewable energy systems.

Applications of Long Short-Term Memory (LSTM) Networks in Polymeric Sciences: A Review.

This review explores the application of Long Short-Term Memory (LSTM) networks, a specialized type of recurrent neural network (RNN), in the field of polymeric sciences. LSTM networks have shown notable effectiveness in modeling sequential data and predicting time-series outcomes, which are essential for understanding complex molecular structures and dynamic processes in polymers. This review delves into the use of LSTM models for predicting polymer properties, monitoring polymerization processes, and evaluating the degradation and mechanical performance of polymers. Additionally, it addresses the challenges related to data availability and interpretability. Through various case studies and comparative analyses, the review demonstrates the effectiveness of LSTM networks in different polymer science applications. Future directions are also discussed, with an emphasis on real-time applications and the need for interdisciplinary collaboration. The goal of this review is to connect advanced machine learning (ML) techniques with polymer science, thereby promoting innovation and improving predictive capabilities in the field.

Experience: A Comparative Analysis of Multivariate Time-Series Generative Models: A Case Study on Human Activity Data

Human activity recognition (HAR) is an active research field that has seen great success in recent years due to advances in sensory data collection methods and activity recognition systems. Deep artificial intelligence (AI) models have contributed to the success of HAR systems lately although still suffering from limitations such as data scarcity, the high costs of labelling data instances and datasets’ imbalance and bias. The temporal nature of human activity data, represented as time series data, impose an additional challenge to using AI models in HAR because most state-of-the-art models do not account for the time component of the data instances. These limitations have inspired the time-series research community to design generative models for sequential data but very little work has been done to evaluate the quality of such models. In this work, we conduct a comparative quality analysis of three generative models for time-series data, using a case study in which we aim to generate sensory human activity data from a seed public dataset. Additionally, we adapt and clearly explain four evaluation methods of synthetic time-series data from the literature and apply them to assess the quality of the synthetic activity data we generate. We show experimentally that high quality human activity data can be generated using deep generative models, and the synthetic data can thus be used in HAR systems to augment real activity data. We also demonstrate that the chosen evaluation methods effectively ensure that the generated data meets the essential quality benchmarks of realism, diversity, coherence and utility. Our findings suggest that using deep generative models to produce synthetic human activity data can potentially address challenges related to data scarcity, biases, and expensive labeling. This holds promise for enhancing the efficiency and reliability of HAR systems.

Multi-Feature Cross Attention-Induced Transformer Network for Hyperspectral and LiDAR Data Classification

Transformers have shown remarkable success in modeling sequential data and capturing intricate patterns over long distances. Their self-attention mechanism allows for efficient parallel processing and scalability, making them well-suited for the high-dimensional data in hyperspectral and LiDAR imagery. However, further research is needed on how to more deeply integrate the features of two modalities in attention mechanisms. In this paper, we propose a novel Multi-Feature Cross Attention-Induced Transformer Network (MCAITN) designed to enhance the classification accuracy of hyperspectral and LiDAR data. The MCAITN integrates the strengths of both data modalities by leveraging a cross-attention mechanism that effectively captures the complementary information between hyperspectral and LiDAR features. By utilizing a transformer-based architecture, the network is capable of learning complex spatial-spectral relationships and long-range dependencies. The cross-attention module facilitates the fusion of multi-source data, improving the network’s ability to discriminate between different land cover types. Extensive experiments conducted on benchmark datasets demonstrate that the MCAITN outperforms state-of-the-art methods in terms of classification accuracy and robustness.

Optimizing Recurrent Neural Networks: A Study on Gradient Normalization of Weights for Enhanced Training Efficiency

Recurrent Neural Networks (RNNs) are classical models for processing sequential data, demonstrating excellent performance in tasks such as natural language processing and time series prediction. However, during the training of RNNs, the issues of vanishing and exploding gradients often arise, significantly impacting the model’s performance and efficiency. In this paper, we investigate why RNNs are more prone to gradient problems compared to other common sequential networks. To address this issue and enhance network performance, we propose a method for gradient normalization of network weights. This method suppresses the occurrence of gradient problems by altering the statistical properties of RNN weights, thereby improving training effectiveness. Additionally, we analyze the impact of weight gradient normalization on the probability-distribution characteristics of model weights and validate the sensitivity of this method to hyperparameters such as learning rate. The experimental results demonstrate that gradient normalization enhances the stability of model training and reduces the frequency of gradient issues. On the Penn Treebank dataset, this method achieves a perplexity level of 110.89, representing an 11.48% improvement over conventional gradient descent methods. For prediction lengths of 24 and 96 on the ETTm1 dataset, Mean Absolute Error (MAE) values of 0.778 and 0.592 are attained, respectively, resulting in 3.00% and 6.77% improvement over conventional gradient descent methods. Moreover, selected subsets of the UCR dataset show an increase in accuracy ranging from 0.4% to 6.0%. The gradient normalization method enhances the ability of RNNs to learn from sequential and causal data, thereby holding significant implications for optimizing the training effectiveness of RNN-based models.

Prediction of Slope Safety Factor Based on Attention Mechanism-Enhanced CNN-GRU

This paper proposes a new method for predicting slope safety factors that combines convolutional neural networks (CNNs), gated recurrent units (GRUs), and attention mechanisms. This method can better capture long-term dependencies, enhance the ability to model sequential data, and reduce the dependence on noisy data, thereby reducing the risk of overfitting. The goal is to improve the accuracy of slope safety factor prediction, detect potential slope stability issues in a timely manner, and take corresponding preventive and control measures to ensure the long-term stability and safety of infrastructure and promote sustainable development. The Pearson correlation coefficient is used to analyze the relationship between the target safety factor and the collected parameters. A one-dimensional CNN layer is used to extract high-dimensional features from the input data, and then a GRU layer is used to capture the correlation between parameters in the sequence. Finally, an attention mechanism is introduced to optimize the weights of the GRU output, enhance the influence of key information, and optimize the overall prediction model. The performance of the proposed model is evaluated using metrics such as the mean absolute error (MAE), mean absolute percentage error (MAPE), mean squared error (MSE), root-mean-square error (RMSE), and R2. The results show that the CNN-GRU-SE model outperforms the GRU, CNN, and CNN-GRU models in terms of prediction accuracy for slope safety factors, with improvements of 4%, 2%, and 1%, respectively. Overall, the research in this paper makes valuable contributions to the field of slope safety factor prediction, and the proposed method also has the potential to be extended to other time-series prediction fields, providing support for a wide range of engineering applications and further promoting the realization of sustainable development.

Multidimensional Forensic Analysis of Onion Sites Using LSTM and Fuzzy Encoded Techniques: Insights and Applications in Digital Investigations

When it comes to digital investigations, the fact that onion sites make people anonymous makes investigative research much harder. For multidimensional forensic analysis of onion sites, this study suggests a new method that combines Long Short-Term Memory (LSTM) networks with fuzzy encoding methods. When you use the Tor network to get to onion sites, both users and services become anonymous, which makes it harder to do regular investigations. LSTM networks are used to look at trends in how people use the site and when material is updated over time. These networks are good at modeling sequential data. Fuzzy encoding methods work with LSTM to handle the uncertainty that comes with analyzing anonymous data, which makes investigative results more accurate. The method involves keeping time logs of what people do on onion sites and turning them into fuzzy sets to show how user interactions and content changes change over time. LSTM models are taught on these recorded sequences to find patterns that point to suspicious activity, like sudden changes in content or strange patterns of user access. This method can be used to find illegal activities, find secret services that host illegal content, and make profiles of how users behave in an anonymous network. Investigators can better track actions across onion sites when they use LSTM's ability to see how events depend on each other over time and fuzzy encoding's ability to deal with inaccurate data. Real-world onion site data used for experiments shows that the proposed method can find secret behaviors and patterns that are hard for traditional investigative methods to find. The results show that the accuracy and efficiency of detecting are much better than with traditional methods.

Cloud-based machine learning algorithms for anomalies detection

Gradient boosting machines harnesses the inherent capabilities of decision trees and meticulously corrects their errors in a sequential fashion, culminating in remarkably precise predictions. Word2Vec, a prominent word embedding technique, occupies a pivotal role in natural language processing (NLP) tasks. Its proficiency lies in capturing intricate semantic relationships among words, thereby facilitating applications such as sentiment analysis, document classification, and machine translation to discern subtle nuances present in textual data. Bayesian networks introduce probabilistic modeling capabilities, predominantly in contexts marked by uncertainty. Their versatile applications encompass risk assessment, fault diagnosis, and recommendation systems. Gated recurrent units (GRU), a variant of recurrent neural networks, emerges as a formidable asset in modeling sequential data. Both training and testing are crucial to the success of an intrusion detection system (IDS). During the training phase, several models are created, each of which can recognize typical from anomalous patterns within a given dataset. To acquire passwords and credit card details, "phishing" usually entails impersonating a trusted company. Predictions of student performance on academic tasks are improved by hyper parameter optimization of the gradient boosting regression tree using the grid search approach.

Deep learning-based modeling of the cyclic behavior of replaceable fuse buckling-restrained braces (BRBs)

This paper presents a novel approach for predicting the behavior of buckling restrained braces (BRBs) using long short-term memory (LSTM) and residual neural network (ResNet) models. Nonlinear finite element analysis (NLFEA) usually requires 3D modeling of complex parts and assigning various material and simulation properties before conducting the analysis, which can be time-consuming and computationally costly. LSTM networks are particularly well-suited for modeling sequential data, such as time series, and ResNet is a discriminative deep learning-based model that can handle deeper architectures without the issue of vanishing gradient. Developing a general LSTM model requires a comprehensive database for training, cross-validation, and blind testing. Even if such a database were available, concatenating data from different BRB specimens would not be feasible when modeling a data series. Such a method would intertwine the features of the individual specimens, rendering the data useless for machine-learning purposes. On the other hand, a well-trained LSTM can predict cyclic behavior for specimens that resemble the training data. As such, a group of well-trained models can produce superior predictions when used on individually familiar data. Hence, there is a need for a prescreening classification stage. The framework presented herein utilizes an AI classifier (ResNet) that automatically recognizes the BRB specimen type of the input data and directs it to the appropriate LSTM model for prediction. The presented framework demonstrates excellent predictions with an accuracy greater than 99%. It is customizable, adaptable, and scalable to include different BRB specimens. The input data consisted of the displacement of the BRB specimens based on the AISC 341–10 qualification recommendations, while the output was the corresponding force (hysteretic response to cyclic loading). The proposed framework can be valuable in earthquake engineering applications as it demonstrates superior predictions.

Deep Learning Backward Recognition Model to Improve Pronunciation Accuracy in the Teaching of Spoken English

The advent of speech recognition algorithms has opened new avenues for enhancing pronunciation accuracy in the teaching of spoken English. As English continues to be a global lingua franca, proficiency in spoken communication holds paramount importance for individuals across diverse professional and academic domains. This paper introduces Forward Backward Recognition Deep Learning (FBRDL), a novel approach aimed at leveraging speech recognition algorithms to enhance pronunciation accuracy in the teaching of spoken English. FBRDL incorporates advanced deep learning techniques such as recurrent neural networks (RNNs) and transformers, which excel in modeling sequential data and capturing long-range dependencies. By leveraging these powerful architectures, FBRDL can effectively handle the inherent variability and complexity of speech signals, enabling robust and accurate recognition even in noisy or adverse environments. Moreover, FBRDL is characterized by its adaptability and scalability, making it well-suited for a wide range of applications across industries. Whether in the realm of virtual assistants, automatic transcription, or voice-controlled devices, FBRDL offers a versatile solution capable of meeting the demands of modern speech recognition tasks. FBRDL integrates principles from deep learning with advanced speech recognition techniques to provide learners with real-time feedback and guidance on their pronunciation. By analyzing spoken English inputs and identifying phonetic discrepancies, FBRDL offers targeted interventions tailored to individual learners' needs. FBRDL achieves an average increase in pronunciation accuracy of 20% compared to traditional teaching methods. Moreover, qualitative assessments underscore the effectiveness of FBRDL in facilitating more precise and efficient acquisition of spoken English skills.

A fault diagnosis method for flexible converter valve equipment based on DSC-BIGRU-MA

Introduction: Precise fault diagnosis is crucial for enhancing the reliability and lifespan of the flexible converter valve equipment. To address this issue, depthwise separable convolution, bidirectional gate recurrent unit, and multi-head attention module (DSC-BiGRU-MAM) based fault diagnosis approach is proposed.Methods: By DSC and BiGRU operation, the model can capture the correlation between local features and temporal information when processing sequence data, thereby enhancing the representation ability and predictive performance of the model for complex sequential data. In addition, by incorporating a multi-head attention module, the proposed method dynamically learns important information from different time intervals and channels. The proposed MAM continuously stimulates fault features in both time and channel dimensions during training, while suppressing fault independent expressions. As a result, it has made an important contribution to improving the performance of the fault diagnosis model.Results and Discussion: Experimental results demonstrate that the proposed method achieves higher accuracy compared to existing methods, with an average accuracy of 95.45%, average precision of 88.67%, and average recall of 89.03%. Additionally, the proposed method has a moderate number of model parameters (17,626) and training time (935 s). Results indicate that the proposed method accurately diagnoses faults in flexible converter valve equipment, especially in real-world situations with noise overlapping signals.

New graphical models for sequential data and the improved state estimations by data-conditioned driving noises

A prevalent problem in statistical signal processing, applied statistics, and time series analysis arises from the attempt to identify the hidden state of Markov process based on a set of available noisy observations. In the context of sequential data, filtering refers to the probability distribution of the underlying Markovian system given the measurements made at or before the time of the estimated state. In addition to the filtering, the smoothing distribution is obtained from incorporating measurements made after the time of the estimated state into the filtered solution. This work proposes a number of new filters and smoothers that, in contrast to the traditional schemes, systematically make use of the process noises to give rise to enhanced performances in addressing the state estimation problem. In doing so, our approaches for the resolution are characterized by the application of the graphical models; the graph-based framework not only provides a unified perspective on the existing filters and smoothers but leads us to design new algorithms in a consistent and comprehensible manner. Moreover, the graph models facilitate the implementation of the suggested algorithms through message passing on the graph.

Quantum Convolutional Long Short-Term Memory Based on Variational Quantum Algorithms in the Era of NISQ

In the era of noisy intermediate-scale quantum (NISQ) computing, the synergistic collaboration between quantum and classical computing models has emerged as a promising solution for tackling complex computational challenges. Long short-term memory (LSTM), as a popular network for modeling sequential data, has been widely acknowledged for its effectiveness. However, with the increasing demand for data and spatial feature extraction, the training cost of LSTM exhibits exponential growth. In this study, we propose the quantum convolutional long short-term memory (QConvLSTM) model. By ingeniously integrating classical convolutional LSTM (ConvLSTM) networks and quantum variational algorithms, we leverage the variational quantum properties and the accelerating characteristics of quantum states to optimize the model training process. Experimental validation demonstrates that, compared to various LSTM variants, our proposed QConvLSTM model outperforms in terms of performance. Additionally, we adopt a hierarchical tree-like circuit design philosophy to enhance the model’s parallel computing capabilities while reducing dependence on quantum bit counts and circuit depth. Moreover, the inherent noise resilience in variational quantum algorithms makes this model more suitable for spatiotemporal sequence modeling tasks on NISQ devices.

Tensor Recurrent Neural Network With Differential Privacy

Recurrent neural network (RNN), a branch of deep learning, is a powerful model for sequential data that has outstanding performance on a wide range of important Internet of Things (IoT) tasks. This unprecedented growth of RNN model has however encountered both heterogeneous IoT data and privacy issues. Existing RNN model can not deal with heterogeneous sequential data; often the larger datasets used in training of RNN model contain sensitive information. To tackle these challenges and for the first time, this research proposes a novel differentially private tensor-based RNN (DPTRNN) that can be applied in many challenging deep learning sequence tasks for IoT systems. Specifically, to process heterogeneous sequential data, we propose a tensor-based RNN model. To guarantee privacy, we develop a tensor-based back-propagation through time algorithm with perturbation to avoid exposing the sensitive information for training the tensor-based RNN model within the framework of differential privacy. Thorough security analysis shows that the differential private tensor-based RNN efficiently protects the confidentiality of sensitive user information for IoT. Our results from extensive experiments on two challenging large video datasets suggest that our proposed scheme is practical with guarantee of data privacy preservation and acceptable accuracy loss.

LiteTransNet: An interpretable approach for landslide displacement prediction using transformer model with attention mechanism

Accurate landslide displacement prediction is crucial for effective early warning systems to mitigate hazards. The importance of historical information varies with time during prediction due to the underlying landslide deformation mechanism. Despite advances in dynamic machine learning models like Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), they struggle to fully capture and interpret the varying importance of historical information during prediction, resulting in decreased accuracy and limited understanding of landslide physical mechanisms. Additionally, these approaches rely on manual feature selection, which is independent of the learning process and therefore is prone to estimation errors. To address these challenges, we propose LiteTransNet, a Lightweighted Transformer Network tailored for landslide displacement prediction. Built on the revolutionary Transformer model for sequential data, LiteTransNet leverages localized self-attention to selectively focus on relevant timestamps and provides interpretable results through its attention heatmap. Furthermore, LiteTransNet performs end-to-end time series modeling without extensive feature engineering. We validate LiteTransNet on two landslides in China's Three Gorges Reservoir Area and demonstrate its improved accuracy over recurrent baselines utilizing feature selection methods. Notably, the attention heatmap generated by LiteTransNet provides interpretable insights into the model's temporal dependency learning. It uncovers the varying importance of historical information at different timestamps, showcasing how LiteTransNet's attention aligns with specific periods of the external environment that cause significant disruptions in the landslide system. Our findings reveal that these disruptions continuously impact displacement prediction until their effects fade or are replaced by subsequent intense disturbances, which traditional recurrent methods are unable to capture. Further experiment shows that LiteTransNet enhances efficiency through its streamlined architecture and its parallelization capability. Overall, LiteTransNet innovates landslide prediction by providing accuracy, interpretability, and efficiency, enhancing understanding of deformation mechanisms to facilitate effective early warning systems.

Optimizing Supply Chain Management Through BO-CNN-LSTM for Demand Forecasting and Inventory Management

This project addresses demand forecasting and inventory optimization in supply chain management. Traditional methods have limitations with complex demand patterns and large-scale data. Deep learning techniques are employed to enhance accuracy and efficiency. The project utilizes BO-CNN-LSTM, leveraging Bayesian optimization for hyperparameter tuning, Convolutional Neural Networks (CNNs) for spatiotemporal feature extraction, and Long Short-Term Memory Networks (LSTMs) for modeling sequential data. Experimental results validate the effectiveness of the approach, outperforming traditional methods. Practical implementation in supply chain management improves operational efficiency and cost control.

Enhanced Intrusion Detection with LSTM-Based Model, Feature Selection, and SMOTE for Imbalanced Data

This study introduces a sophisticated intrusion detection system (IDS) that has been specifically developed for internet of things (IoT) networks. By utilizing the capabilities of long short-term memory (LSTM), a deep learning model renowned for its proficiency in modeling sequential data, our intrusion detection system (IDS) effectively discerns between regular network traffic and potential malicious attacks. In order to tackle the issue of imbalanced data, which is a prevalent concern in the development of intrusion detection systems (IDSs), we have integrated the synthetic minority over-sampling technique (SMOTE) into our approach. This incorporation allows our model to accurately identify infrequent incursion patterns. The rebalancing of the dataset is accomplished by SMOTE through the generation of synthetic samples belonging to the minority class. Various strategies, such as the utilization of generative adversarial networks (GANs), have been put forth in order to tackle the issue of data imbalance. However, SMOTE (synthetic minority over-sampling technique) presents some distinct advantages when applied to intrusion detection. The SMOTE is characterized by its simplicity and proven efficacy across diverse areas, including in intrusion detection. The implementation of this approach is straightforward and does not necessitate intricate adversarial training techniques such as generative adversarial networks (GANs). The interpretability of SMOTE lies in its ability to generate synthetic samples that are aligned with the properties of the original data, rendering it well suited for security applications that prioritize transparency. The utilization of SMOTE has been widely embraced in the field of intrusion detection research, demonstrating its effectiveness in augmenting the detection capacities of intrusion detection systems (IDSs) in internet of things (IoT) networks and reducing the consequences of class imbalance. This study conducted a thorough assessment of three commonly utilized public datasets, namely, CICIDS2017, NSL-KDD, and UNSW-NB15. The findings indicate that our LSTM-based intrusion detection system (IDS), in conjunction with the implementation of SMOTE to address data imbalance, outperforms existing methodologies in accurately detecting network intrusions. The findings of this study provide significant contributions to the domain of internet of things (IoT) security, presenting a proactive and adaptable approach to safeguarding against advanced cyberattacks. Through the utilization of LSTM-based deep learning techniques and the mitigation of data imbalance using SMOTE, our AI-driven intrusion detection system (IDS) enhances the security of internet of things (IoT) networks, hence facilitating the wider implementation of IoT technologies across many industries.

Diagnostics and predictive maintenance of industrial electric motors

Automatic continuous or periodic control is the most promising method of diagnosing electric motors in productionnowadays. Itis aimed at predicting breakdowns and the remaining useful lifetimeof motors. However, research in this arearemains purely theoretical and eitherfocuses on very narrow problemsor providestoo superficial overview.Consequently,manual control devices orpartially automated devices are mainlyusedin practice[1]. In view of this, there is great interest in the development of a software method for predictive maintenance of electric motors. In this research recurrent neural networks with long short-term memory layers (LSTM-layers)are investigated due to their abilityto effectively model sequential data and learn complex dependencies [2].

Calibration-free blood pressure estimation based on a convolutional neural network.

In this study, we conducted research on a deep learning-based blood pressure (BP) estimation model suitable for wearable environments. To measure BP while wearing a wearable watch, it needs to be considered that computing power for signal processing is limited and the input signals are subject to noise interference. Therefore, we employed a convolutional neural network (CNN) as the BP estimation model and utilized time-series electrocardiogram (ECG) and photoplethysmogram (PPG) signals, which are quantifiable in a wearable context. We generated periodic input signals and used differential and thresholding methods to decrease noise in the preprocessing step. We then applied a max-pooling technique with filter sizes of 2 × 1 and 5 × 1 within a 3-layer convolutional neural network to estimate BP. Our method was trained, validated, and tested using 2.4 million data samples from 49 patients in the intensive care unit. These samples, totaling 3.1 GB were obtained from the publicly accessible MIMIC database. As a result of a test with 480,000 data samples, the average root mean square error in BP estimation was 3.41, 5.80, and 2.78 mm Hg in the prediction of pulse pressure, systolic BP (SBP), and diastolic BP (DBP), respectively. The cumulative error percentage less than 5 mm Hg was 68% and 93% for SBP and DBP, respectively. In addition, the cumulative error percentage less than 15 mm Hg was 98% and 99% for SBP and DBP. Subsequently, we evaluated the impact of changes in input signal length (1 cycle vs. 30 s) and the introduction of noise on BP estimation results. The experimental results revealed that the length of the input signal did not significantly affect the performance of CNN-based analysis. When estimating BP using noise-added ECG signals, the mean absolute error (MAE) for SBP and DBP was 9.72 and 6.67 mm Hg, respectively. Meanwhile, when using noise-added PPG signals, the MAE for SBP and DBP was 26.85 and 14.00 mm Hg, respectively. Therefore, this study confirmed that using ECG signals rather than PPG signals is advantageous for noise reduction in a wearable environment. Besides, short sampling frames without calibration can be effective as input signals. Furthermore, it demonstrated that using a model suitable for information extraction rather than a specialized deep learning model for sequential data can yield satisfactory results in BP estimation.