Term Time Series Data Research Articles

Using a VAE-SOM architecture for anomaly detection of flexible sensors in limb prosthesis

Flexible wearable sensor electronics, combined with advanced software functions, pave the way toward increasingly intelligent healthcare devices. One important application area is limb prosthesis, where printed flexible sensor solutions enable efficient monitoring and assessing of the actual intra-socket dynamic operation conditions in clinical and other more natural environments. However, the data collected by such sensors suffer from variations and errors, leading to difficulty in perceiving the actual operational conditions. This paper proposes a novel method for detecting anomalies in the data that are collected for measuring the intra-socket dynamic operation conditions by printed flexible wearable sensors. A discrete generative model based on Variational AutoEncoder (VAE) is used first to encode the collected multi-variant time-series data in terms of latent states. After that, a clustering method based on the Self-Organizing Map (SOM) is used to acquire discrete and interpretable representations of the VAE encoded latent states. An adaptive Markov chain is utilized to detect anomalies by quantifying state transitions and revealing temporal dependencies. The contributions of the proposed architecture conclude as follows: (1) Using the VAE-SOM hybrid model to regularize the continues data as discrete states, supporting interpreting the operational data to analytic models. (2) Employing adaptive Markov chains to generalize the transitions of these states, allowing to model the complex operational conditions. Compared with benchmark methods, our architecture is validated via two public datasets and achieves the best F1 scores. Moreover, we measure the run-time performance of this lightweight architecture. The results indicate that the proposed method performs low computational complexity, facilitating the applications on real-life productions.

Analysis of Total Factor Productivity of Power and Energy Sector in Bangladesh: A Study on CSE Listed Companies

Data envelopment analysis (DEA) is one of the most prominent productivity analysis model which measures overall performance of a firm based on specified input and output factors considering time series data. Power and energy sector of Bangladesh are controlled by both state owned and institutional ownership. The aim of the study is to assess Productivity performance of privately owned power and energy sector of Bangladesh by considering salaries and operating expenditure as input factors and revenue profit and total asset are output factors of eleven companies listed under Chittagong Stock Exchange (CSE) of Bangladesh. Malmquist output oriented productivity analysis was used to examine six years’ time series data, finds that total factor productivity of power and energy sector under private ownership is increased by 2.5% indicating firms overall efficiency is increased within sample period of time. The limitation of study is that long term time series data and state owned Organization are not considered in this study. Further research may be carried out considering both state owned and privatized institutions as well as more input-output variables may be used to examine the effectiveness of this sector on the whole economy.

Long term very low frequency ambient noise: A window to the ocean and atmosphere

Long term time series data of ambient noise in deep water from a site near Wake Island in the western Pacific provide a window into processes occurring in the ocean and in the atmosphere. Acoustic noise data in the Very Low and Ultra Low Frequency bands show strong frequency components at major tidal frequencies, are correlated with El Nino Sea Surface Temperature Anomaly time records, and correlate also with yearly wind fluctuations. Directional ambient noise data also allow tracking of distant storms. Noise at 1 Hz was chosen in particular for study because the noise sources are primarily environmental and, thus, the data may serve as a harbinger of pending environmental change.

Trends Analysis of Stream Flow at Different Gauging Stations in Upper Jhelum River

Analysis of stream flow trends plays a significant role in the flood forecasting and hydrologic drought assessment related investigations besides study of water resource management In this study, the long-term stream flow data recorded at different gauging stations in Upper Jhelum River (UJR) in Jammu and Kashmir (J &K) state basin were estimated using the Modified Mann-Kendall Test (MMKT) and Sen’s Slope Estimator (SSE).Long term time series data of ten gauging stations used for trend analysis were annual maximum and monthly maximum stream flows for the years (1980-2016) for the months of occurrence of flood in UJRB. It was observed that six gauging stations out of the ten showed a significant decreasing trend in discharge and depth of flow in the UJRB at 0.05 probability level of significance. However the results of the monthly and seasonal maximum stream flow analysis showed that during spring season (March to May) about 90% of the gauging stations depicted downward trends in the summer season (Jun to Aug) all the gauging stations showed downward stream flow trends and approximately 60% of the gauging stations showed downward trends in the month of September (first month of autumn). This analysis resulted into quantification of flood magnitudes and its periodicity of occurrences to develop flood mitigation strategies. Nonetheless, such investigation of water flow in the river would also assist in regional hydrologic studies besides occurrence of flood and hydrologic droughts.

Dynamic interaction between tidal current and upstream discharge at the Kapuas River mouth

This paper explores the dynamic interaction between tidal current and upstream discharge at the Kapuas River mouth. Long term time series data of tidal current and river discharge obtained from software Windwave 05 and hydrograph data, respectively. Three prominent parameters, i.e., river discharge, tides, and tidal current were analyzed using the wavelet method, Hilbert function, and low pass filter to evaluate the tidal characteristics and phase difference of M2, M4, K1, O1. The relationship between the tidal current and water level shows a mimic standing wave with the phase difference around ∼ -180 – 900, indicating the tidal current leads the water level. The low-pass filter results indicate that the increasing river discharge hampers the tides and reduces the tidal current at the mouth. The Hilbert function result reveals the asymmetric tide and flood dominant flow which has the opposite direction with the seaward direction of the current. Moreover, the subtidal water level flows in a seaward direction. This happened because the net seaward current modulated by the river discharge. The currents at the mouth are affected significantly by the swell wave.

APPLICATION OF THE ECONOMETRIC APPROACH TO FORECASTING THE KEY INDICATORS IN RENTAL OPERATIONS

Introduction. Lease is considered one of the alternative forms of financing and a means of providing production and economic activities with the necessary assets. Leasing operations are relevant and cost-effective for small and medium-sized businesses, which do not always have enough resources to buy their own premises, equipment, etc. Leases have many characteristics and can be described using quantitative research methods. Methods. The practical application of scientific methods in our study is as follows: dialectical - to determine general approaches to research; monographic - in relation to the in-depth study of certain social phenomena; scientific knowledge - to determine a certain "perspective" of research as an assumption about the most relevant to the study of the selected problem characteristics of the object; abstract-logical - in order to return from the formulated scientific abstractions again to the study of the concrete and the formulation of the necessary conditions for its further existence and development in a logical sequence; scientific generalization - regarding the generalization of scientific facts and results based on the study, as well as methods of statistical processing of information, using methods of obtaining such quantities and qualitative characteristics, which are not directly in the economic information of the enterprise under study. Results. The article analyzes the influence of various factors on the amount of rent using correlation and regression analysis. The structure of leasing agreements and lease transactions in Ukraine by the volume of sold services is analyzed. Leasing operations in Ukraine by the volume of realized services in terms of real estate lease and the amount of rent in the studied lease agreements are considered. A linear model based on a spatial series of data is built, and its quality is checked. The initial model was modified taking into account the information about the statistical insignificance of individual variables. The forecast of the rent indicator is made taking into account the set factor indicators. Discussion. A promising area of further research is to improve the model by analyzing additional array of information (search for additional significant factors), as well as the use of econometric tools to analyze time series data in terms of lease and use them for short-term forecasting. Keywords: lease (rent), leasing, rental, correlation and regression analysis, model, forecasting.

Fractal Analysis of Black Carbon in the Coal Mine Regions of India

Time series display irregular patterns and self-similar microstructures usually termed as fractals over a period of time, depending on several factors like climate, season, market fluctuations, growth cycles etc. Various methods are developed in mathematical modeling to investigate such trends in the time series. Black carbon is one of the major air pollutants released during several mining activities having dangerous effect on climate and human health. In this paper we aim to study the behavioral pattern and trends in the time series of black carbon concentration over the three major coal mines of India namely Raniganj, Jharia and Bokaro considering a long term time series data of past 38 years using regression and fractal dimensional analysis of black carbon between the major coal mines with the help statistical tools like Hurst exponent, predictability index and trend percent. In this comparative study it is observed that the time series display anti persistent behavior with positive and large variation then the past time. The results and comparisons obtained will be helpful in studying the behavior and trends in the concentration of black carbon over the coal mines regions. The significance of the study will be helpful to gather the interest of researcher’s, NGO’s and government towards the black carbon concentration over the coal mines so that preventive measures and new policies be framed to keep a check on black carbon concentration in these areas.

Long-term Performance Analysis of CdTe PV module in real operating conditions

Solar photovoltaic (PV) technology has gained ample attention with the continuously increasing electricity demand and climate concerns. The Cadmium Telluride (CdTe) based PV technology is the leading and foremost technology among rapidly growing thin film PV industry because of its lower energy and economic payback period and energy efficiency competitive to the wafer based technology. Although, these modules undergo a series of indoor qualification tests specified by IEC61646, the long-term performance analysis in real operation conditions is essential in order to ensure long term stability, robustness and reliability. In this paper, performance analysis of 80WP CdTe PV module has been carried out on the basis of long term time series data of short circuit current (ISC) and open circuit voltage (VOC) measured in outdoor conditions. The measured electrical parameters has been translated to alternate reporting conditions i.e. 800W/m2 irradiance over the module surface and 40°C module temperature. The monthly averaged and normalized VOC and ISC of the module is found to be degraded by 2.86% and 2.29% after 23 months of outdoor exposure. The parameters are translated to standard test conditions for diurnal analysis which shows the degradation of 2.91% for VOC while no deviation in the value of ISC.

Autoencoder for wind power prediction

Successful integration of renewable energy sources like wind power into smart grids largely depends on accurate prediction of power from these intermittent sources. Production of wind power cannot be controlled as the wind speed can vary based on weather conditions. Accurate prediction of wind power can assist smart grid that intelligently decides on the usage of alternative power sources based on demand forecast. Time series wind speed data are normally used for wind power prediction. In this paper, we have investigated the usage of a set of secondary features obtained using deep learning for wind power prediction. Deep learning is a special form on neural network that is capable of capturing the structural properties of time series data in terms of a set of numeric features. More precisely, we have designed a two-stage autoencoder (a particular type of deep learning) and incorporated the structural features into a prediction framework. Using the structural features, we have achieved as high as 12.63% better prediction accuracy than traditionally used statistical features.

Visualizing State of Time-Series Data by Supervised Gaussian Process Dynamical Models

Gaussian Process Dynamical Models (GPDMs) constitute a nonlinear dimensionality reduction technique that provides a probabilistic representation of time series data in terms of Gaussian process priors. In this paper, we report a method based on GPDMs to visualize the states of time-series data. Conventional GPDMs are unsupervised, and therefore, even when the labels of data are available, it is not possible to use this information. To overcome the problem, we propose a supervised GPDM (S-GPDM) that utilizes both the data and their corresponding labels. We demonstrate experimentally that the S-GPDM can locate related motion data closer together than conventional GPDMs.

An Improved Mathematical Model for Computing Power Output of Solar Photovoltaic Modules

It is difficult to determine the input parameters values for equivalent circuit models of photovoltaic modules through analytical methods. Thus, the previous researchers preferred to use numerical methods. Since, the numerical methods are time consuming and need long term time series data which is not available in most developing countries, an improved mathematical model was formulated by combination of analytical and numerical methods to overcome the limitations of existing methods. The values of required model input parameters were computed analytically. The expression for output current of photovoltaic module was determined explicitly by Lambert W function and voltage was determined numerically by Newton-Raphson method. Moreover, the algebraic equations were derived for the shape factor which involves the ideality factor and the series resistance of a single diode photovoltaic module power output model. The formulated model results were validated with rated power output of a photovoltaic module provided by manufacturers using local meteorological data, which gave ±2% error. It was found that the proposed model is more practical in terms of precise estimations of photovoltaic module power output for any required location and number of variables used.

Electricity Generation Technologies Systematic Risk

We propose a method to estimate differentiated costs of equity for electricity generation technologies. We provide evidence that there is a substantial difference in their cost of equity. This method is based on a regression between available information on electric utilities equity return and their electricity generation asset portfolio. Our first main finding is that there is a statistically significant risk premium for nuclear technology. This point is in line with the most recent studies using a higher discount rate for nuclear technology compared to coal fired plants. But, in contrast with these studies, we find that the value of the risk premium is on the order of 1% above the systematic risk of a coal-fired plant, which is considerably lower than their recommended value of 3%. The robustness of this result is tested for the structural breaks that may have been induced by the August, 2008 financial crisis and the Daishi-Fukushima accident of March, 2011. Our second finding deals with renewable energies. Despite the existence of subsidies through feed-in tariffs which would make them look like safe investments, they present a high level of systematic risk. A possible explanation for this important excess return is the fact that they are dependent on subsidies for their development and thus, bear a larger systematic exposure than can be seen on short term time series data.

Energy Time Series Data Analysis based on a Novel Integrated Data Characteristic Testing Approach

This paper attempts to propose an integrated data characteristic testing approach for energy time series data so as to analyze the energy dynamics, which serves as the foundation for the model selection problem. Based on thoroughly analyzing the main data characteristics of energy time series data together with their interrelationship, these data characteristics are divided into two main categories: nature and pattern characteristics to explore energy time series data from different perspectives. In nature determination, the energy time series data is analyzed in terms of nonstationarity, nonlinearity and complexity characteristics from a global perspective. In pattern measurements, the characteristics of cyclicity (and seasonality), mutability (or saltation) and randomicity (or noise pattern) signify the relative hidden patterns and the impacts on the original data, via a way of decomposition. For illustration purpose, hydropower consumptions in China and USA are analyzed and the main data characteristics are thoroughly explored by using the proposed integrated approach. Empirical results reveal that besides same characteristics of difference strationarity, nonlinearity and seasonality, the hydropower markets in China and USA are quite different: while China's hydropower market are comparatively simple but sensitive to emergencies, e.g., government support and technological progress, US’ hydropower market is otherwise mature and efficient with the nature of high leveled complexity and the main pattern of randomicity. The results also confirm the proposed integrated approach an effective tool to test energy time series data in terms of data characteristics, paving the way for the further model formulation and forecasting.

Stability of the nonlinear dynamics of an optically injected VCSEL

Automated protocols have been developed to characterize time series data in terms of stability. These techniques are applied to the output power time series of an optically injected vertical cavity surface emitting laser (VCSEL) subject to varying injection strength and optical frequency detuning between master and slave lasers. Dynamic maps, generated from high resolution, computer controlled experiments, identify regions of dynamic instability in the parameter space.

Delayed Effects of Climate Variables on Incidence of Dengue in Singapore during 2000-2010

In this paper we apply time-series analysis to study the association between dengue incidence and climate variables such as temperature, humidity and precipitation while taking into account the delayed effects and further discussing the differences between using short term data (each year) and long term data from 2000 to 2010. Time-lag correlation method is employed to determine the time delay between incidence of dengue and climate variables. The result is not only consistent with previous results of other researchers, it also provides new findings. After analyzing the data in each year (short term time series data) and 10-year data (long term time series data), We find that using the short term data and the long term data can provide complementary insights into the relationship between dengue incidence and climate variables.

Delayed Effects of Weather Variables on Incidence of Dengue Fever in Singapore from 2000-2010

In this paper we analyze the time-series data to study the association between dengue incidence and weather variables such as temperature and humidity while taking into account the delayed effects and further discussing the differences between using short term data (yearly) and long term data in Singapore from 2000 to 2010. The time-series analysis and statistical analysis method are employed to determine the delay in timing between the incidence of dengue fever and weather variables. The results are not only consistent with previous results of other researchers, but also provide new findings. We analyze the data from each year (short term time series data) and 10-year data (long term time series data) and find that the use of both the short term data and the long term data can provide complementary insights into the relationship between dengue incidence and weather variables.

Wavelet Frequency Domain Approach for Statistical Modeling of Rainfall Time-Series Data

The powerful methodology of “Wavelet analysis in frequency domain” for analyzing time-series data is studied. As an illustration, Indian monsoon rainfall time-series data from 1879–2006 is considered. The entire data analysis is carried out using SPLUS WAVELET TOOLKIT software package. The discrete wavelet transform (DWT) and multiresolution analysis (MRA) of the data are computed to analyze the behaviour of trend present in the time-series data in terms of different times and scales. By using bootstrap method, size and power of the test for testing significance of trend in the data is computed. It is found that the size of the test for Daubechies wavelet is more than that for Haar wavelet. In respect of both Daubechies and Haar wavelet filters, it is found that the test for presence of trend is unbiased. Also, power of the test for both Daubechies (D4) and Haar wavelets, at level 5 is less than the one at level 6. Further, Haar wavelet at level 6 has generally performed better than Daubechies (D4) wavelet at level 6 in terms of power of the test. Using the former wavelet, a declining trend in the data under consideration is revealed.

A fuzzy logic algorithm to assign confidence levels to heart and respiratory rate time series

We have developed a fuzzy logic-based algorithm to qualify the reliability of heart rate (HR) and respiratory rate (RR) vital-sign time-series data by assigning a confidence level to the data points while they are measured as a continuous data stream. The algorithm's membership functions are derived from physiology-based performance limits and mass-assignment-based data-driven characteristics of the signals. The assigned confidence levels are based on the reliability of each HR and RR measurement as well as the relationship between them. The algorithm was tested on HR and RR data collected from subjects undertaking a range of physical activities, and it showed acceptable performance in detecting four types of faults that result in low-confidence data points (receiver operating characteristic areas under the curve ranged from 0.67 (SD 0.04) to 0.83 (SD 0.03), mean and standard deviation (SD) over all faults). The algorithm is sensitive to noise in the raw HR and RR data and will flag many data points as low confidence if the data are noisy; prior processing of the data to reduce noise allows identification of only the most substantial faults. Depending on how HR and RR data are processed, the algorithm can be applied as a tool to evaluate sensor performance or to qualify HR and RR time-series data in terms of their reliability before use in automated decision-assist systems.