Time Series Classification Tasks Research Articles

AsMODiTS: An application of surrogate models to optimize Time Series symbolic discretization through archive-based training set update strategy

The enhanced multi-objective symbolic discretization for time series (eMODiTS) uses an evolutionary process to identify the appropriate discretization scheme in the Time Series Classification (TSC) task. It discretizes using a unique alphabet cut for each word segment. However, this kind of scheme has a higher computational cost. Therefore, this study implemented surrogate models to minimize this cost. The general procedure is summarized below.•The K-nearest neighbor for regression, the support vector regression model, and the Ra- dial Basis Functions neural networks were implemented as surrogate models to estimate the objective values of eMODiTS, including the discretization process.•An archive-based update strategy was introduced to maintain diversity in the training set.•Finally, the model update process uses a hybrid (fixed and dynamic) approach for the surrogate model's evolution control.

Towards the generalization of time series classification: A feature-level style transfer and multi-source transfer learning perspective

Transfer learning-based methods hold promise for enhancing classification task performance. However, a transfer learning mechanism for hard-to-classify time series classification tasks caused by limited samples in training set is still to be accomplished. Drawing inspiration from domain adaptation and style transfer methods in computer vision, we aim to address sample scarcity and enhance classifier generalization by leveraging abundant unrelated time series datasets. In this paper, a transfer learning mechanism called feature-level style transfer, where the feature alignment is conducted in a domain adaptation-inspired manner, has been proposed to target initially hard-to-classify time series tasks with limited training samples. The proposed mechanism eliminates constraints on source dataset selection, enabling the utilization of weakly-related or unrelated datasets to enhance classifier performance. Furthermore, a voting mechanism has been formulated to achieve more accurate predictions by taking transferable information learnt from different source domains into comprehensive consideration.

Multivariate time series classification with crucial timestamps guidance

Transformer-based deep learning methods have significantly facilitated multivariate time series classification (MTSC) tasks. However, due to the inherent operation of self-attention mechanism, most existing methods tend to overlook the internal local features and temporal invariance of time series, potentially resulting in a limited understanding of the representation and context information within the model. In contrast to global features, local features demonstrate greater specificity and detail, thereby being more conducive to capture essential texture information and local structures of time series. To ameliorate these problems, we propose CTNet, a novel network that enhances the time series representation learning by reconstructing C̲rucial T̲imestamps, aiming to improve the ability to address MTSC tasks. Specifically, we introduce a novel Transformer encoder that incorporates a highly effective Gaussian-prior mechanism to accurately capture local dependencies. Additionally, we present a data-driven mask strategy to boost model’s capability of representation learning by reconstructing crucial timestamps. During the reconstruction process, we employ context-aware positional encoding to augment the temporal invariance of the model. Extensive experiments conducted on 30 accessible UEA datasets validate the superiority of CTNet compared to previous competitive methods. Furthermore, ablation studies and visualization analyses are conducted to confirm the effectiveness of the proposed model.

Augmentation blending with clustering-aware outlier factor: An outlier-driven perspective for enhanced contrastive learning

Data augmentation is an effective strategy to improve model performance and generalization capabilities by expanding both the size and diversity of the original dataset via subtle modifications to the original data. In mainstream contrastive learning frameworks, data augmentation is leveraged by default to compose positive and negative pairs. However, selecting optimal augmentations that consistently preserve the semantics of the original samples poses challenges owing to variations in the quality of different augmentation transformations. To address this challenge, we provide a novel perspective and formulate the data augmentation optimization problem as an outlier-driven augmentation blending problem. Specifically, we design an innovative metric named Clustering-Aware Outlier Factor (CAOF), to quantify the semantic inconsistency between the augmented samples and original samples. CAOF is capable of identifying semantically inconsistent outliers in augmented samples that are generated by various transformations, thereby offering better context-aware density assessments by considering the intra-cluster neighbors for individual samples. Each augmentation transformation is evaluated using the proposed CAOF, which reflects the likelihood of producing semantically inconsistent augmented samples. The likelihood determines the importance of the transformation in generating a refined blended augmentation, which is subsequently fed into the contrastive learning framework. We conduct empirical evaluations on six time series datasets. We compare our method with three state-of-the-art augmentation selection methods on five contrastive learning models. The results demonstrate the superiority of our proposed approach in discovering outliers and refining augmentations in contrastive learning for time series classification tasks.

Time Series Classification Based on Forward Echo State Convolution Network

The Echo state network (ESN) is an efficient recurrent neural network that has achieved good results in time series prediction tasks. Still, its application in time series classification tasks has yet to develop fully. In this study, we work on the time series classification problem based on echo state networks. We propose a new framework called forward echo state convolutional network (FESCN). It consists of two parts, the encoder and the decoder, where the encoder part is composed of a forward topology echo state network (FT-ESN), and the decoder part mainly consists of a convolutional layer and a max-pooling layer. We apply the proposed network framework to the univariate time series dataset UCR and compare it with six traditional methods and four neural network models. The experimental findings demonstrate that FESCN outperforms other methods in terms of overall classification accuracy. Additionally, we investigated the impact of reservoir size on network performance and observed that the optimal classification results were obtained when the reservoir size was set to 32. Finally, we investigated the performance of the network under noise interference, and the results show that FESCN has a more stable network performance compared to EMN (echo memory network).

Time and frequency-domain feature fusion network for multivariate time series classification

Multivariate time series classification is a significant research topic in the realm of data mining, which encompasses a wide array of practical applications in domains such as healthcare, energy systems, and traffic. The complex temporal and spatial dependencies inherent in multivariate time series pose challenges to classification tasks. Previous studies have usually focused on the time-domain information of multivariate time series. However, achieving accurate classification using only the time-domain information may be difficult. To address this challenge, a time and frequency-domain feature fusion network (TF-Net) for multivariate time series classification is proposed in this paper. Our model contains two modules, the time-domain module and the frequency-domain module. The time-domain module is used to capture the time-domain features of multivariate time series. It is constructed using CNNs and GCNs, enabling the capture of both temporal and spatial dependencies within the time-domain. The frequency-domain module is used to capture the frequency-domain features of multivariate time series data. In this module, we treat the frequency-domain features as images and innovatively transform the multivariate time series classification task into an image classification task. Our method is able to classify multivariate time series from both the time-domain and frequency-domain, which provides a new perspective for multivariate time series analysis. We conduct extensive experiments on the UAE archive, and the experimental results show that our model achieves the best performance compared to the ten state-of-the-art methods.

Safety control strategy of spinal lamina cutting based on force and cutting depth signals

AbstractLaminectomy is one of the most common posterior spinal operations. Since the lamina is adjacent to important tissues such as nerves, once damaged, it can cause serious complications and even lead to paralysis. In order to prevent the above injuries and complications, ultrasonic bone scalpel and surgical robots have been introduced into spinal laminectomy, and many scholars have studied the recognition method of the bone tissue status. Currently, almost all methods to achieve recognition of bone tissue are based on sensor signals collected by high‐precision sensors installed at the end of surgical robots. However, the previous methods could not accurately identify the state of spinal bone tissue. Innovatively, the identification of bone tissue status was regarded as a time series classification task, and the classification algorithm LSTM‐FCN was used to process fusion signals composed of force and cutting depth signals, thus achieving an accurate classification of the lamina bone tissue status. In addition, it was verified that the accuracy of the proposed method could reach 98.85% in identifying the state of porcine spinal laminectomy. And the maximum penetration distance can be controlled within 0.6 mm, which is safe and can be used in practice.

Predicting intubation for intensive care units patients: A deep learning approach to improve patient management

ObjectiveFor patients in the Intensive Care Unit (ICU), the timing of intubation has a significant association with patients' outcomes. However, accurate prediction of the timing of intubation remains an unsolved challenge due to the noisy, sparse, heterogeneous, and unbalanced nature of ICU data. In this study, our objective is to develop a workflow for pre-processing ICU data and to develop a customized deep learning model to predict the need for intubation. MethodsTo improve the prediction accuracy, we transform the intubation prediction task into a time series classification task. We carefully design a sequence of data pre-processing steps to handle the multimodal noisy data. Firstly, we discretize the sequential data and address missing data using interpolation. Next, we employ a sampling strategy to address data imbalance and standardize the data to facilitate faster model convergence. Furthermore, we employ the feature selection technique and propose an ensemble model to combine features learned by different deep learning models. ResultsThe performance is evaluated on Medical Information Mart for Intensive Care (MIMIC)-III, an ICU dataset. Our proposed Deep Feature Fusion method achieves an area under the curve (AUC) of the receiver operating curve (ROC) of 0.8953, surpassing the performance of other deep learning and traditional machine learning models. ConclusionOur proposed Deep Feature Fusion method proves to be a viable approach for predicting intubation and outperforms other deep learning and classical machine learning models. The study confirms that high-frequency time-varying indicators, particularly Mean Blood Pressure (MeanBP) and peripheral oxygen saturation (SpO2), are significant risk factors for predicting intubation.

Diffusion Language-Shapelets for Semi-supervised Time-Series Classification

Semi-supervised time-series classification could effectively alleviate the issue of lacking labeled data. However, existing approaches usually ignore model interpretability, making it difficult for humans to understand the principles behind the predictions of a model. Shapelets are a set of discriminative subsequences that show high interpretability in time series classification tasks. Shapelet learning-based methods have demonstrated promising classification performance. Unfortunately, without enough labeled data, the shapelets learned by existing methods are often poorly discriminative, and even dissimilar to any subsequence of the original time series. To address this issue, we propose the Diffusion Language-Shapelets model (DiffShape) for semi-supervised time series classification. In DiffShape, a self-supervised diffusion learning mechanism is designed, which uses real subsequences as a condition. This helps to increase the similarity between the learned shapelets and real subsequences by using a large amount of unlabeled data. Furthermore, we introduce a contrastive language-shapelets learning strategy that improves the discriminability of the learned shapelets by incorporating the natural language descriptions of the time series. Experiments have been conducted on the UCR time series archive, and the results reveal that the proposed DiffShape method achieves state-of-the-art performance and exhibits superior interpretability over baselines.

DSDCLNet: Dual-stream encoder and dual-level contrastive learning network for supervised multivariate time series classification

In recent years, deep learning approaches have shown remarkable advancements in Multivariate Time Series Classification (MTSC) tasks. However, the existing approaches primarily focus on capturing the long-term correlations of time series or identifying local key sequence fragments, inevitably neglecting the synergistic properties between global and local components. Additionally, most representation learning methods for MTSC rely on self-supervised learning, which limits their ability to fully exploit label information. Hence, this paper proposes a novel approach termed Dual-Stream Encoder and Dual-Level Contrastive Learning Network (DSDCLNet), which integrates a Dual-Stream Encoder (DSE) and Dual-level Contrastive Learning (DCL). First, to extract multiscale local-global features from multivariate time series data, we employ a DSE architecture comprising an Attention-Gated Recurrent Unit (AGRU) and a Dual-layer Multiscale Convolutional Neural Network (DMSCNN). Specifically, DMSCNN consists of a series of multi-scale convolutional layers and a max pooling layer. Second, to maximize the utilization of label information, a new loss function is designed, which combines classification loss, instance-level contrastive loss, and temporal-level contrastive loss. Finally, experiments are conducted on the UEA datasets and the results demonstrate that DSDCLNet achieves the highest average accuracy of 0.77, outperforming traditional approaches, deep learning approaches, and self-supervised approaches on 30, 23, and 27 datasets, respectively.

Universal representation learning for multivariate time series using the instance-level and cluster-level supervised contrastive learning

AbstractThe multivariate time series classification (MTSC) task aims to predict a class label for a given time series. Recently, modern deep learning-based approaches have achieved promising performance over traditional methods for MTSC tasks. The success of these approaches relies on access to the massive amount of labeled data (i.e., annotating or assigning tags to each sample that shows its corresponding category). However, obtaining a massive amount of labeled data is usually very time-consuming and expensive in many real-world applications such as medicine, because it requires domain experts’ knowledge to annotate data. Insufficient labeled data prevents these models from learning discriminative features, resulting in poor margins that reduce generalization performance. To address this challenge, we propose a novel approach: supervised contrastive learning for time series classification (SupCon-TSC). This approach improves the classification performance by learning the discriminative low-dimensional representations of multivariate time series, and its end-to-end structure allows for interpretable outcomes. It is based on supervised contrastive (SupCon) loss to learn the inherent structure of multivariate time series. First, two separate augmentation families, including strong and weak augmentation methods, are utilized to generate augmented data for the source and target networks, respectively. Second, we propose the instance-level, and cluster-level SupCon learning approaches to capture contextual information to learn the discriminative and universal representation for multivariate time series datasets. In the instance-level SupCon learning approach, for each given anchor instance that comes from the source network, the low-variance output encodings from the target network are sampled as positive and negative instances based on their labels. However, the cluster-level approach is performed between each instance and cluster centers among batches, as opposed to the instance-level approach. The cluster-level SupCon loss attempts to maximize the similarities between each instance and cluster centers among batches. We tested this novel approach on two small cardiopulmonary exercise testing (CPET) datasets and the real-world UEA Multivariate time series archive. The results of the SupCon-TSC model on CPET datasets indicate its capability to learn more discriminative features than existing approaches in situations where the size of the dataset is small. Moreover, the results on the UEA archive show that training a classifier on top of the universal representation features learned by our proposed method outperforms the state-of-the-art approaches.

TSCF: An Improved Deep Forest Model for Time Series Classification

The deep forest presents a novel approach that yields competitive performance when compared to deep neural networks. Nevertheless, there are limited studies on the application of deep forest to time series classification (TSC) tasks, and the direct use of deep forest cannot effectively capture the relevant characteristics of time series. For that, this paper proposes time series cascade forest (TSCF), a model specifically designed for TSC tasks. TSCF relies on four base classifiers, i.e., random forest, completely random forest, random shapelet forest, and diverse representation canonical interval forest, allowing for feature learning on the original data from three granularities: point, subsequence, and summary statistics calculated based on intervals. The major contribution of this work, is to define an ensemble and deep classifier that significantly outperforms the individual classifiers and the original deep forest. Experimental results show that TSCF outperforms other forest-based algorithms for solving TSC problems.

IMU-Based Fitness Activity Recognition Using CNNs for Time Series Classification.

Mobile fitness applications provide the opportunity to show users real-time feedback on their current fitness activity. For such applications, it is essential to accurately track the user's current fitness activity using available mobile sensors, such as inertial measurement units (IMUs). Convolutional neural networks (CNNs) have been shown to produce strong results in different time series classification tasks, including the recognition of daily living activities. However, fitness activities can present unique challenges to the human activity recognition task (HAR), including greater similarity between individual activities and fewer available data for model training. In this paper, we evaluate the applicability of CNNs to the fitness activity recognition task (FAR) using IMU data and determine the impact of input data size and sensor count on performance. For this purpose, we adapted three existing CNN architectures to the FAR task and designed a fourth CNN variant, which we call the scaling fully convolutional network (Scaling-FCN). We designed a preprocessing pipeline and recorded a running exercise data set with 20 participants, in which we evaluated the respective recognition performances of the four networks, comparing them with three traditional machine learning (ML) methods commonly used in HAR. Although CNN architectures achieve at least 94% test accuracy in all scenarios, two traditional ML architectures surpass them in the default scenario, with support vector machines (SVMs) achieving 99.00 ± 0.34% test accuracy. The removal of all sensors except one foot sensor reduced the performance of traditional ML architectures but improved the performance of CNN architectures on our data set, with our Scaling-FCN reaching the highest accuracy of 99.86 ± 0.11% on the test set. Our results suggest that CNNs are generally well suited for fitness activity recognition, and noticeable performance improvements can be achieved if sensors are dropped selectively, although traditional ML architectures can still compete with or even surpass CNNs when favorable input data are utilized.

Accelerating Neural ODEs Using Model Order Reduction.

Embedding nonlinear dynamical systems into artificial neural networks is a powerful new formalism for machine learning. By parameterizing ordinary differential equations (ODEs) as neural network layers, these Neural ODEs are memory-efficient to train, process time series naturally, and incorporate knowledge of physical systems into deep learning (DL) models. However, the practical applications of Neural ODEs are limited due to long inference times because the outputs of the embedded ODE layers are computed numerically with differential equation solvers that can be computationally demanding. Here, we show that mathematical model order reduction (MOR) methods can be used for compressing and accelerating Neural ODEs by accurately simulating the continuous nonlinear dynamics in low-dimensional subspaces. We implement our novel compression method by developing Neural ODEs that integrate the necessary subspace-projection and interpolation operations as layers of the neural network. We validate our approach by comparing it to neuron pruning and singular value decomposition (SVD)-based weight truncation methods from the literature in image and time-series classification tasks. The methods are evaluated by acceleration versus accuracy when adjusting the level of compression. On this spectrum, we achieve a favorable balance over existing methods by using MOR when compressing a convolutional Neural ODE. In compressing a recurrent Neural ODE, SVD-based weight truncation yields good performance. Based on our results, our integration of MOR with Neural ODEs can facilitate efficient, dynamical system-driven DL in resource-constrained applications.

Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects.

Self-supervised learning (SSL) has recently achieved impressive performance on various time series tasks. The most prominent advantage of SSL is that it reduces the dependence on labeled data. Based on the pre-training and fine-tuning strategy, even a small amount of labeled data can achieve high performance. Compared with many published self-supervised surveys on computer vision and natural language processing, a comprehensive survey for time series SSL is still missing. To fill this gap, we review current state-of-the-art SSL methods for time series data in this article. To this end, we first comprehensively review existing surveys related to SSL and time series, and then provide a new taxonomy of existing time series SSL methods by summarizing them from three perspectives: generative-based, contrastive-based, and adversarial-based. These methods are further divided into ten subcategories with detailed reviews and discussions about their key intuitions, main frameworks, advantages and disadvantages. To facilitate the experiments and validation of time series SSL methods, we also summarize datasets commonly used in time series forecasting, classification, anomaly detection, and clustering tasks. Finally, we present the future directions of SSL for time series analysis.

Time series classification models based on nonlinear spiking neural P systems

Reservoir computing (RC) is a novel class of recurrent neural networks (RNN) models. Nonlinear spiking neural P (NSNP) systems are neural-like computing models with nonlinear spiking mechanisms. By introducing NSNP systems as the reservoir, we propose a new RC model for time series classification task, termed TSC-NSNP model. However, due to the high-dimensional nature of the reservoir state space, the TSC-NSNP model, like existing RC models, will encounter some challenges. To address the challenges. we utilize the reservoir model space representation and dimensionality reduction method to propose two improved models, termed TSC-DR-NSNP model and TSC-RMS-NSNP model. The three RC models can be easily realized and learnt in the RC framework. The proposed three RC models are evaluated on 21 benchmark time series classification data sets, and are compared with 20 classification models. The comparisons demonstrate the effectiveness of the presented three RC models for time series classification tasks.

Multi-Scale Ensemble Booster for Improving Existing TSD Classifiers

Time Series Classification (TSC) is an essential task in Time Series Data (TSD) analysis. Ensemble-based approaches now achieve the best performance on TSC tasks. However, integrating numerous different models makes them highly suffer from heavy preprocessing. Even worse, non-deep-learning ensemble-based methods suffer from substantial computational costs due to lacking GPU acceleration. Multi-scale information in TSD can improve TSC performance. However, Existing TSD classifiers employing multi-scale information struggle with heavy preprocessing and cannot help other TSD classifiers obtain multi-scale feature extraction capabilities. Inspired by these, we proposed a performance enhancement framework called multi-scale ensemble booster (MEB), helping existing TSD classifiers achieve performance leaps. In MEB, we proposed an easy-to-combine network structure without changing any of their structure and hyperparameters, only needed to set one hyperparameter, consisting of multi-scale transformation and multi-output decision fusion. Then, a probability distribution co-evolution strategy is proposed to attain the optimal label probability distribution. We conducted numerous ablation experiments of MEB on 128 univariate datasets and 29 multivariate datasets and comparative experiments with 11 state-of-the-art methods, which demonstrated the significant performance improvement ability of MEB and the most advanced performance of the model enhanced by MEB, respectively. Furthermore, to figure out why MEB can improve model performance, we provided a chain of interpretability analyses.

Reservoir computing models based on spiking neural P systems for time series classification

Nonlinear spiking neural P (NSNP) systems are neural-like membrane computing models with nonlinear spiking mechanisms. Because of this nonlinear spiking mechanism, NSNP systems can show rich nonlinear dynamics. Reservoir computing (RC) is a novel recurrent neural network (RNN) and can overcome some shortcomings of traditional RNNs. Based on NSNP systems, we developed two RC variants for time series classification, RC-SNP and RC-RMS-SNP, which are without and integrated with reservoir model space (RMS), respectively. The two RC variants use NSNP systems as the reservoirs and can be easily implemented in the RC framework. The proposed two RC variants were evaluated on 17 benchmark time series classification datasets and compared with 16 state-of-the-art or baseline classification models. The comparison results demonstrate the effectiveness of the proposed two RC variants for time series classification tasks.

Multi-scale self-supervised representation learning with temporal alignment for multi-rate time series modeling

Deep sequential networks have shown great power in time series regression and classification. So far, most approaches naturally assume that the time sequential data are uniformly sampled. In practice, however, different variables usually have different sampling rates, thereby forming multi-rate time series (MR-TS). Particularly, the target variable (i.e., label) to be predicted usually has a lower sampling frequency due to the difficulty of manual annotations. The multi-rate problem poses two challenges. One is the diverse dynamics at different sampling rates, which is defined as multi-scale dynamics. The other is label scarcity. To tackle the above obstacles, this paper developed a Multi-scale Self-supervised Representation Learning technique with a Temporal Alignment mechanism (MSRL-TA) as a coherent framework. Concretely, a probabilistic masked autoencoding approach is pertinently developed, in which segment-wise masking schemes and rate-aware positional encodings are devised to enable the characterization of multi-scale temporal dynamics. In the course of pre-training, the encoder networks are able to generate rich and holistic representations of multi-rate data, thereby alleviating the label scarcity issue for supervised fine-tuning. Furthermore, a Temporal Alignment mechanism is devised to refine synthesized features for dynamic predictive modeling through feature block division and block-wise convolution. With empirical evaluation through extensive experiments, our proposed MSRL-TA achieved consistent state-of-the-art in both multi-rate time series regression and classification tasks on five real-world datasets, including air quality prediction, industrial soft sensing, human activity recognition, and speech voice classification.

CALIMERA: A new early time series classification method

Early time series classification is a variant of the time series classification task, in which a label must be assigned to the incoming time series as quickly as possible without necessarily screening through the whole sequence. It needs to be realized on the algorithmic level by fusing a decision-making method that detects the right moment to stop and a classifier that assigns a class label. The contribution addressed in this paper is twofold. Firstly, we present a new method for finding the best moment to perform an action (terminate/continue). Secondly, we propose a new learning scheme using classifier calibration to estimate classification accuracy. The new approach, called CALIMERA, is formalized as a cost minimization problem. Using two benchmark methodologies for early time series classification, we have shown that the proposed model achieves better results than the current state-of-the-art. Two most serious competitors of CALIMERA are ECONOMY and TEASER. The empirical comparison showed that the new method achieved a higher accuracy than TEASER for 35 out of 45 datasets and it outperformed ECONOMY in 20 out of 34 datasets.