Large Amount Of Time Series Data Research Articles

DLT-GAN: Dual-Layer Transfer Generative Adversarial Network-Based Time Series Data Augmentation Method

In actual production processes, analysis and prediction tasks commonly rely on large amounts of time-series data. However, real-world scenarios often face issues such as insufficient or imbalanced data, severely impacting the accuracy of analysis and predictions. To address this challenge, this paper proposes a dual-layer transfer model based on Generative Adversarial Networks (GANs) aiming to enhance the training speed and generation quality of time-series data augmentation under small-sample conditions while reducing the reliance on large training datasets. This method introduces a module transfer strategy based on the traditional GAN framework which balances the training between the discriminator and the generator, thereby improving the model’s performance and convergence speed. By employing a dual-layer network structure to transfer the features of time-series signals, the model effectively reduces the generation of noise and other irrelevant features, improving the similarity of the generated signals’ characteristics. This paper uses speech signals as a case study, addressing scenarios where speech data are difficult to collect and the limited number of speech samples available for effective feature extraction and analysis. Simulated speech timbre generation is conducted, and the experimental results on the CMU-ARCTIC database show that, compared to traditional methods, this approach achieves significant improvements in enhancing the consistency of generated signal features and the model’s convergence speed.

Enhancing EEG data quality and precision for cloud-based clinical applications: an evaluation of the SLOG framework

Automation is revamping our preprocessing pipelines, and accelerating the delivery of personalized digital medicine. It improves efficiency, reduces costs, and allows clinicians to treat patients without significant delays. However, the influx of multimodal data highlights the need to protect sensitive information, such as clinical data, and safeguard data fidelity. One of the neuroimaging modalities that produces large amounts of time-series data is Electroencephalography (EEG). It captures the neural dynamics in a task or resting brain state with high temporal resolution. EEG electrodes placed on the scalp acquire electrical activity from the brain. These electrical potentials attenuate as they cross multiple layers of brain tissue and fluid yielding relatively weaker signals than noise-low signal-to-noise ratio. EEG signals are further distorted by internal physiological artifacts, such as eye movements (EOG) or heartbeat (ECG), and external noise, such as line noise (50 Hz). EOG artifacts, due to their proximity to the frontal brain regions, are particularly challenging to eliminate. Therefore, a widely used EOG rejection method, independent component analysis (ICA), demands manual inspection of the marked EOG components before they are rejected from the EEG data. We underscore the inaccuracy of automatized ICA rejection and provide an auxiliary algorithm-Second Layer Inspection for EOG (SLOG) in the clinical environment. SLOG based on spatial and temporal patterns of eye movements, re-examines the already marked EOG artifacts and confirms no EEG-related activity is mistakenly eliminated in this artifact rejection step. SLOG achieved a 99% precision rate on the simulated dataset while 85% precision on the real EEG dataset. One of the primary considerations for cloud-based applications is operational costs, including computing power. Algorithms like SLOG allow us to maintain data fidelity and precision without overloading the cloud platforms and maxing out our budgets.

Evaluation of data imputation approaches for multi-stream building systems data1

Increasing advancements in building digitization, smart sensing, and metering technologies have allowed large amounts of timeseries data to be collected for monitoring, analyzing, and controlling building systems. However, due to sensor or communication failures, the data collected are often incomplete and poor in quality. Data imputation approaches to replace the missing values, specifically based on either statistical or predictive models have been widely adopted for multivariate datasets in other domains. It is hence of interest to find an effective way to impute timeseries data collected from a building system. In this paper, we evaluate multiple data imputation approaches using data collected from a medium sized building situated in Stockholm, Sweden and a small commercial building from the ASHRAE RP-1312 research project. Sensors with widely varying characteristics from the case study buildings were selected to evaluate the imputation methods. The imputation accuracy and the impact of each chosen imputation method on information entropy, short-term building forecasting model performance, and fault detection strategy were evaluated. Results demonstrate that incorporating time-lagged cross correlations within a k -nearest neighbor ( k NN) model provide the most accurate imputations without affecting the quality of subsequent data analysis.

Temporal profiling of sediment microbial communities in the Three Gorges Reservoir Area discovered time-dissimilarity patterns and multiple stable states

Microbial communities play vital roles in cycling nutrients and maintaining water quality in aquatic ecosystems. To better understand the dynamics of microbial communities and to pave way to effective ecological remediation, it's essential to reveal the temporal patterns of the communities and to identify their states. However, research exploring the dynamic changes of microbial communities needs a large amount of time-series data, which could be an extravagant requirement for a single study. In this research, we overcame this challenge by conducting a meta-analysis of years of accumulations of 16S rRNA high-throughput sequencing data from the Three Gorges Reservoir Area (TGRA), an ecological and environmental hotspot. For better understanding the microbial communities time-dissimilarity dynamics, three microbial communities time-dissimilarity patterns were hypothesized, and the linear pattern in the TGRA was validated. In addition, to explore the stability of microbial communities in the TGRA, two alternative stable states were revealed, and their differences in community richness, alpha diversity indices, community composition, ecological network topological properties, and metabolic functions were demonstrated. In short, two states of microbial communities showed distinct richness and alpha diversity indices, and the communities in one state were more dominated by Halomonas and Nitrosopumilaceae genera, facilitating nitrogen cycling metabolic processes; whilst the main genera of the other state were Bathyarchaeia and Methanosaeta, which favored methane-related metabolism. Moreover, different studies and environmental differences between mainstream and tributaries were attributed as the potential inducing factors of the state division. Our study provides a comprehensive insight into the dynamics and stability of microbial communities in the TGRA, and a reference for future studies on microbial community dynamics.

Building Analytics Tool Deployment at Scale: Benefits, Costs, and Deployment Practices

Buildings are becoming more data-rich. Building analytics tools, including energy information systems (EIS) and fault detection and diagnostic (FDD) tools, have emerged to enable building operators to translate large amounts of time-series data into actionable findings to achieve energy and non-energy benefits. To expedite data analytics adoption and facilitate technology innovation, building owners, technology developers, and researchers need reliable cost–benefit data and evidence-based guidance on deployment practices. This paper fulfills these needs with the energy use and survey data from a wide-ranging research and industry partnership program that covers thousands of buildings installed with analytics tools. The paper indicates that after two years of implementation, organizations using FDD tools and EIS tools achieved 9% and 3% median annual energy savings, respectively. The median base cost and annual recurring cost for FDD are USD 0.65 per square meter (m2) (USD 0.06 per square foot [ft2]) and USD 0.22 per m2 (USD 0.02 per ft2), and are USD 0.11 per m2 (USD 0.01 per ft2) and USD 0.11 per m2 (USD 0.01 per ft2) for EIS. The common metrics and analyses that are used in the tools to support the discovery of energy efficiency measures are summarized in detail. Two best practice examples identified to maximize the benefits of tool implementation are also presented. Opportunities to advance the state of technology include simplified data integration and management, and more efficient processes for acting on analytics outputs. Compared with previous efforts in the literature, the findings presented in this paper demonstrate the effectiveness of building analytics tools with the largest known dataset.

Support Vector Machine and Granular Computing Based Time Series Volatility Prediction

With the development of information technology, a large amount of time-series data is generated and stored in the field of economic management, and the potential and valuable knowledge and information in the data can be mined to support management and decision-making activities by using data mining algorithms. In this paper, three different time-series information granulation methods are proposed for time-series information granulation from both time axis and theoretical domain: time-series time-axis information granulation method based on fluctuation point and time-series time-axis information granulation method based on cloud model and fuzzy time-series prediction method based on theoretical domain information granulation. At the same time, the granulation idea of grain computing is introduced into time-series analysis, and the original high-dimensional time series is granulated into low-dimensional grain time series by information granulation of time series, and the constructed information grains can portray and reflect the structural characteristics of the original time-series data, to realize efficient dimensionality reduction and lay the foundation for the subsequent data mining work. Finally, the grains of the decision tree are analyzed, and different support vector machine classifiers corresponding to each grain are designed to construct a global multiclassification model.

An efficient data collection framework in the sky: An affine transformation approach based on Internet of unmanned aerial vehicles

Due to a large amount of time-series data (e.g. precipitation, temperature, humidity) in the air, reliable and fast data collection is a challenging issue. Using an unmanned aerial vehicle (UAV) as a data collector to collect time-series data is an effective method. However, due to the limited storage capacity of the UAV, the UAV must access the sensor multiple times to collect time-series data from the deployed sensors and dump the data to the data centre, which leads to significantly increased time and cost for data collection mission. To address this challenge, an efficient time-series data collection framework is proposed based on the Internet of UAVs. In this system, a min-maximum data processing strategy is adopted based on data value to store the collected time-series data. Specifically, efficient data compression storage is achieved with minimal loss of precision by extracting the dominant dataset with maximum value. Furthermore, an efficient affine transformation method is proposed to improve the efficiency of the system. Extensive case studies on some real-world datasets demonstrate that the proposed framework can achieve efficient data management and compressed storage.

Deep Transfer Learning for Time Series Data Based on Sensor Modality Classification.

The scarcity of labelled time-series data can hinder a proper training of deep learning models. This is especially relevant for the growing field of ubiquitous computing, where data coming from wearable devices have to be analysed using pattern recognition techniques to provide meaningful applications. To address this problem, we propose a transfer learning method based on attributing sensor modality labels to a large amount of time-series data collected from various application fields. Using these data, our method firstly trains a Deep Neural Network (DNN) that can learn general characteristics of time-series data, then transfers it to another DNN designed to solve a specific target problem. In addition, we propose a general architecture that can adapt the transferred DNN regardless of the sensors used in the target field making our approach in particular suitable for multichannel data. We test our method for two ubiquitous computing problems—Human Activity Recognition (HAR) and Emotion Recognition (ER)—and compare it a baseline training the DNN without using transfer learning. For HAR, we also introduce a new dataset, Cognitive Village-MSBand (CogAge), which contains data for 61 atomic activities acquired from three wearable devices (smartphone, smartwatch, and smartglasses). Our results show that our transfer learning approach outperforms the baseline for both HAR and ER.

Simple and Quick Visualization of Periodical Data Using Microsoft Excel.

Actograms are well-established methods used for visualizing periodic activity of animals in chronobiological research. They help in the understanding of the overall characteristics of rhythms and are instrumental in defining the direction of subsequent detailed analysis. Although there exists specialized software for creating actograms, new users such as students and researchers from other fields often find it inconvenient to use. In this study, we demonstrate a fast and easy method to create actograms using Microsoft Excel. As operations in Excel are simple and user-friendly, it takes only a few minutes to create an actogram. Using this method, it is possible to obtain a visual understanding of the characteristics of rhythms not only from typical activity data, but also from any kind of time-series data such as body temperature, blood sugar level, gene expressions, sleep electroencephalogram, heartbeat, and so on. The actogram thus created can also be converted to the "heatogram” shown by color temperature. As opposed to conventional chronograms, this new type of chronogram facilitates easy understanding of rhythmic features in a more intuitive manner. This method is therefore convenient and beneficial for a broad range of researchers including students as it aids in the better understanding of periodic phenomena from a large amount of time-series data.

TimeCluster: dimension reduction applied to temporal data for visual analytics

There is a need for solutions which assist users to understand long time-series data by observing its changes over time, finding repeated patterns, detecting outliers, and effectively labeling data instances. Although these tasks are quite distinct and are usually tackled separately, we present an interactive visual analytics system and approach that can address these issues in a single system. It enables users to visualize, understand and explore univariate or multivariate long time-series data in one image using a connected scatter plot. It supports interactive analysis and exploration for pattern discovery and outlier detection. Different dimensionality reduction techniques are used and compared in our system. Because of its power of extracting features, deep learning is used for multivariate time-series along with 2D reduction techniques for rapid and easy interpretation and interaction with large amount of time-series data. We deploy our system with different time-series datasets and report two real-world case studies that are used to evaluate our system.

Visualizing Multidimensional Lifelogging Data: A Case Study on MyMovieHistory Project

ABSTRACTA large amount of time-series data has been frequently used to extract the useful patterns and trends and to visualize them for better understanding. This work is focusing on visualizing personal lifelogging data for tracking back to personal histories. Thereby, we present several similarity measures between multidimensional data at two different time points. For human evaluation, the method has been applied to MyMovieHistory (which a social recommendation system by storing personal movie logs) and tested with many users. Experimental results shown that the proposed visualization method and interfaces can help to understand user history.

Data-based method for creating electricity use load profiles using large amount of customer-specific hourly measured electricity use data

The recent technological developments monitoring the electricity use of small customers provides with a whole new view to develop electricity distribution systems, customer-specific services and to increase energy efficiency. The analysis of customer load profile and load estimation is an important and popular area of electricity distribution technology and management. In this paper, we present an efficient methodology, based on self-organizing maps (SOM) and clustering methods (K-means and hierarchical clustering), capable of handling large amounts of time-series data in the context of electricity load management research. The proposed methodology was applied on a dataset consisting of hourly measured electricity use data, for 3989 small customers located in Northern-Savo, Finland. Information for the hourly electricity use, for a large numbers of small customers, has been made available only recently. Therefore, this paper presents the first results of making use of these data. The individual customers were classified into user groups based on their electricity use profile. On this basis, new, data-based load curves were calculated for each of these user groups. The new user groups as well as the new-estimated load curves were compared with the existing ones, which were calculated by the electricity company, on the basis of a customer classification scheme and their annual demand for electricity. The index of agreement statistics were used to quantify the agreement between the estimated and observed electricity use. The results indicate that there is a clear improvement when using data-based estimations, while the new-estimated load curves can be utilized directly by existing electricity power systems for more accurate load estimates.

Fractal Analysis of Time-Series Rule-Based Models and Nonlinear Model Predictive Control

Abundant time-series dynamic data can be accumulated from a chemical plant during long-term operations. In our previous work, these plant data were directly implemented for the purpose of model predictive control. However, a large amount of time-series data is required to perform high-quality nonlinear model predictive control. In this work, fractal analysis is performed to reduce the size of a time-series data set for high-quality nonlinear model predictive control. Results in this study indicate that on-line identification of nonlinear models is unnecessary if the disturbances to the process satisfy the fractal-equivalence condition. Simulation examples, including the dual composition control of a high-purity distillation column, demonstrate that the nonlinear model predictive scheme is quite useful for those cases in which the linear model predictive controller has failed.