Clustering Of Time Series Data Research Articles

RETRACTED ARTICLE: ANN and fuzzy based household energy consumption prediction with high accuracy

Timeline Data is gathered according to different time intervals, which are day after day after week or month after month, for updating properties and rationing institutional resilience it is important to consider the usage of systems and lead to reduced lifespan Such details illustrate the use of the system as well as its interaction with time, like day, week, month and time of year, and the relation between the equipment and a relative, essential factors for the effects of the usage of their potency and the expected movement by customers. This is because it is not significant to determine the various relations between specific devices utilizing concurrent data. In addition, precise relations between interval-based instances in which specific system usage continue for certain duration cannot be calculated. To address these difficulties, we propose supervised energy time series data clustering and frequent pattern mining analysis as well as a Bayesian network forecast for energy use. However, the AI model is a univariate construct based on past use-values. Neural Networks have the favored position that can estimate nonlinear limits. Everything together they have an approximate usage of vitality, the ANN adds in a planning knowledge table between the use of vitality (EC) and its determinants. SVM is capable of reliably calculating knowledge on time structure while the basic system mechanism is frequently nonlinear and not set. Also, certain nonlinear mechanisms such as multilayer perceptron have been shown to flank SVM. The single data has been converted into a multivariate and the ANFIS has been selected as it transmits both the AI (ANN) and Fuzzy Inference Method (FIS) points of concern. ANFIS yields the accuracy, RMSE, and MAPE among genuine and anticipated power utilization of 91.19%, 0.4076 and 0.9049 which is moderately low.

Clustering of large scale QoS time series data in federated clouds using improved variable Chromosome Length Genetic Algorithm (CQGA)

Service monitoring in federated clouds generates large scale QoS time series data with various unknown, frequent and abnormal patterns. This could be associated with inaccurate resource provisioning and avoid violations through predictive and preventive actions. A sufficient intelligence in the form of expert system for decision support is needed in such situations. Therefore, the main challenge here is to efficiently discover unknown frequent and abnormal patterns from QoS time series data of federated clouds. On the other hand, QoS time series data in federated clouds is unlabeled and consists of frequent and abnormal structures. Studies showed that clustering is the most common and efficient method to discover interesting patterns and structures from unlabeled data. But, clustering is normally associated with time overhead that should be optimized as well as accuracy issues mainly in connection with convergence and finding an optimum number of clusters. This work proposes a new genetic based clustering algorithm that shows better accuracy and speed in comparison to state-of-the-art methods. Furthermore, the proposed algorithm can find the optimum number of clusters concurrently with the clustering itself. Achieved accuracy and convergence of the proposed method in the experimental results assure its use in expert systems, mainly for resource provisioning and further autonomous decision making situations in federated clouds. In addition to the scientific impact of this paper, the proposed method can be used by federated cloud service providers in practice.

Trend-based time series data clustering for wind speed forecasting

Wind forecasting is a time series problem, can aide in estimating the annual energy production of potential wind farms. Seasonality and trend are the two significant components that characterize the wind time series data. Variability in trend and seasonal component affects the performance of most of the forecasting methods. Therefore, to simplify the wind forecasting technique, generally, nonlinear seasonal and trend components are eliminated from wind time series data. Accuracy depends on the application function that is applicable to eliminate the trend and seasonality. In this article, a hybrid approach for time series forecasting has been proposed. A clustering technique has been developed, which finds the clusters of time series data showing identical trend components. After finding the proper clusters of similar trend components, statistical methods, namely, autoregressive integrated moving average and generalized autoregressive score techniques, are applied to the individual cluster. In the end, resulting components are aggregated. The experiment shows that the cluster-based forecasting technique gives better performance as compared with existing statistical models.

Simulator Pre-Screening of Underprepared Drivers Prior to Licensing On-Road Examination: Clustering of Virtual Driving Test Time Series Data.

BackgroundA large Midwestern state commissioned a virtual driving test (VDT) to assess driving skills preparedness before the on-road examination (ORE). Since July 2017, a pilot deployment of the VDT in state licensing centers (VDT pilot) has collected both VDT and ORE data from new license applicants with the aim of creating a scoring algorithm that could predict those who were underprepared.ObjectiveLeveraging data collected from the VDT pilot, this study aimed to develop and conduct an initial evaluation of a novel machine learning (ML)–based classifier using limited domain knowledge and minimal feature engineering to reliably predict applicant pass/fail on the ORE. Such methods, if proven useful, could be applicable to the classification of other time series data collected within medical and other settings.MethodsWe analyzed an initial dataset that comprised 4308 drivers who completed both the VDT and the ORE, in which 1096 (25.4%) drivers went on to fail the ORE. We studied 2 different approaches to constructing feature sets to use as input to ML algorithms: the standard method of reducing the time series data to a set of manually defined variables that summarize driving behavior and a novel approach using time series clustering. We then fed these representations into different ML algorithms to compare their ability to predict a driver’s ORE outcome (pass/fail).ResultsThe new method using time series clustering performed similarly compared with the standard method in terms of overall accuracy for predicting pass or fail outcome (76.1% vs 76.2%) and area under the curve (0.656 vs 0.682). However, the time series clustering slightly outperformed the standard method in differentially predicting failure on the ORE. The novel clustering method yielded a risk ratio for failure of 3.07 (95% CI 2.75-3.43), whereas the standard variables method yielded a risk ratio for failure of 2.68 (95% CI 2.41-2.99). In addition, the time series clustering method with logistic regression produced the lowest ratio of false alarms (those who were predicted to fail but went on to pass the ORE; 27.2%).ConclusionsOur results provide initial evidence that the clustering method is useful for feature construction in classification tasks involving time series data when resources are limited to create multiple, domain-relevant variables.

An autoencoder-based deep learning approach for clustering time series data

This paper introduces a two-stage deep learning-based methodology for clustering time series data. First, a novel technique is introduced to utilize the characteristics (e.g., volatility) of the given time series data in order to create labels and thus enable transformation of the problem from an unsupervised into a supervised learning. Second, an autoencoder-based deep learning model is built to model both known and hidden non-linear features of time series data. The paper reports a case study in which the selected financial and stock time series data of over 70 stock indices are clustered into distinct groups using the introduced two-stage procedure. The results show that the proposed methodology is capable of achieving 87.5% accuracy in clustering and predicting the labels for unseen time series data. The paper also reports an important finding in which it is observed that the performance of both techniques (i.e., autoencoder and Kmeans) are comparable. However, there are a few instances of time series data that are classified differently by the autoencoder-based methodology compared to the Kmeans algorithm. The results may indicate that the proposed deep learning-based approach is taking into account additional hidden features that might be overlooked by conventional Kmeans. The finding raises the question whether the explicit features of data should be analyzed for clustering or more advanced techniques such as deep learning need to be adapted by which hidden features and relationships are explored for clustering purposes.

Fuzzy clustering based on feature weights for multivariate time series

As an important set of techniques for data mining, time series clustering methods had been studied by many researchers. Although most existing solutions largely focus on univariate time series clustering, there has been a surge in interest in the clustering of multivariate time series data. In this paper, a feature-weighted clustering method is proposed based on two distance measurement methods called dynamic time warping (DTW) and shape-based distance (SDB). There are four stages in the proposed clustering algorithm. First, we pick cluster centers by the pop clustering method called clustering by fast search and find of density peaks (DPC). Next, by considering the overall matching of multivariate time series, a fuzzy membership matrix is generated by performing DTW on all variables. We then reconsider the contribution of each independent dimension by utilizing SBD to measure distances within each dimension and construct multiple fuzzy membership matrices. Finally, we utilize a traditional fuzzy clustering algorithm called fuzzy c-means to cluster the fuzzy membership matrices and generate clustering results. Simultaneously, a feature weight calculation method and novel equation for constructing fuzzy membership matrices are applied during the clustering process. We compare the proposed method to other clustering methods and the results indicate that the proposed method can improve clustering accuracy for multivariate time series datasets.

Operating mode recognition of iron ore sintering process based on the clustering of time series data

Operators often make different control decisions for different operating modes to meet the production requirement of the iron ore sintering process. Recognizing the operating modes is important to improve the quality and quantity of the sinter ore. An operating mode recognition method based on the clustering of time series data for the iron ore sintering process is presented in this paper. First, the Spearman rank correlation analysis and the information entropy analysis are combined to select parameters. Next, the operating mode recognition submodel is built by the fuzzy C-Means clustering method based on dynamic time warping distance and the naive Bayesian classifier method. Then, the outputs of the submodels are fused to obtain the final recognized operating mode. Finally, the productivity and combustion efficiency are regarded as the classification criteria, and the raw data collected from an iron and steel plant are used for the experiment. The experimental results show that the proposed method can effectively recognize the operating mode of the sintering process.

A new approach to clustering time series data using the ARIMA model uncertainty

The piccolo distance is a simplified approach in clustering time-series data, and it needs the data analysts to determine the ARIMA model for each series. Some problems may arise in the modeling step because different criteria may lead to different orders of the best models. This current paper is discussing how to handle this model uncertainty problem by borrowing the concept of an ensemble in the area of data science. Instead of using a single criterion to identify the best model, we proposed to generate the best models using several different criteria. Each series was characterized by the average of the estimates of model parameters obtained. In the clustering process, we employed a hierarchical approached where the optimal number of clusters identified using the Silhouette coefficient. An extensive simulation was completed within this research, and we revealed that our proposed methodology could increase the correct cluster membership by more than 10%. We also implemented our methodology to identify clusters of areas in Indonesia (i.e., Province of Banten) based on the pattern of rainfall level and found an impressive result. Keywords: model-based clustering; uncertainty model; time series clustering; rainfall data.

Application of machine learning methods to assess progressive cavity pumps (PCPs) performance in coal seam gas (CSG) wells

In Queensland, progressive cavity pumps (PCPs) are the artificial lift method of choice in coal seam gas (CSG) wells, and this choice of artificial lift production stems from the ability of PCPs to better manage the production of liquids with suspended solids. As with any mechanical pumping system, PCPs are prone to natural wear and tear over their operational life, and with the production of coal fines and inter-burden, the run life of PCPs in CSG wells is significantly reduced. Another factor to consider with the use of PCPs is their reliability. As per the CSG production data available through the Queensland Government Data Portal, there are approximately 6400 wells operational in the state as of December 2018. This number is expected to grow significantly over the next decade to meet both international and domestic gas utilisation requirements. Operators supervising these wells rely on a reactive or exception-based approach to manage well performance. In order to efficiently operate thousands of PCP wells, it is pertinent that a benchmark methodology is devised to autonomously monitor PCP performance and allow operators to manage wells by exception. In this study, we will cover the application of machine learning methods to understand anomalous PCP behaviour and overall pump performance based on the analysis of multivariate time-series data. An innovative time-series data approximation and image conversion technique will be discussed in this paper, along with machine learning methods, which will focus on a scalable and autonomous approach to cluster PCP performance and detection of anomalous pump behaviour in near real-time. Results from this study show that clustering real-time data based on converted time-series images helps to pro-actively detect change in PCP performance. Discovery of anomalous multivariate events is also achieved through time-series image conversion. This study also demonstrates that clustering time-series data noticeably improves the real-time monitoring capabilities of PCP performance through improved visual analytics.

Trendlets: A novel probabilistic representational structures for clustering the time series data

Time series data is a sequence of values recorded systematically over a period which are mostly used for prediction, clustering, and analysis. The two essential features of a time series data are trend and seasonality. Preprocessing of the time series data is necessary for performing prediction tasks. In most of the cases, the trend and the seasonality are removed before applying the regression algorithms. The accuracy of such algorithms depends upon the functions used for the removal of trend and seasonality. Clustering of an unlabeled time series data with the presence of trend and seasonality is challenging. In this paper, we propose a probabilistic representational learning method for grouping the time series data. We introduce five terminologies in our method of clustering namely the trendlets, uplets, downlets, equalets and trendlet string. These elements are the representational building blocks of our proposed method. Experiments on the proposed algorithm are performed with the renewable energy data on the electricity supply system of continental Europe which includes the demand and inflow of renewable energy for the term 2012 to 2014 and UCR-2018 time series archive containing 128 datasets. We compared our proposed representational method with various clustering algorithms using the silhouette score. Mini-batch k-means and agglomerative hierarchical clustering algorithms show better performance in terms of quality, logical accordance with data and time taken for clustering.

Modified Dynamic Time Warping for Hierarchical Clustering

Time series clustering is the process of grouping sequential correspondences in similar clusters. The key feature behind clustering time series data lies on the similarity/distance function used to identify the sequential matches. Dynamic Time Warping (DTW) is one of the common distance measures that have demonstrated competitive results compared to other functions. DTW aims to find the shortest path in the process of identifying sequential matches. DTW relies on dynamic programming to obtain the shortest path where the smaller distance is being computed. However, in the case of equivalent distances, DTW is selecting the path randomly. Hence, the selection could be misguided in such randomization process, which significantly affects the matching quality. This is due to randomization may lead to the longer path which drifts from obtaining the optimum path. This paper proposes a modified DTW that aims to enhance the dynamic selection of the shortest path when handling equivalent distances. Experiments were conducted using twenty UCR benchmark datasets. Also, the proposed modified DTW result has been compared with the state of the art competitive distance measures which is based on precision, recall and f-measure including the original DTW, Minkowski distance measure and Euclidean distance measure. The results showed that the proposed modified DTW reveal superior results in compared to the standard DTW, either using Minkowski or Euclidean. This can demonstrate the effectiveness of the proposed modification in which optimizing the shortest path has enhanced the performance of clustering. The proposed modified DTW can be used for having good clustering method for any time series data.

Soybean Price Pattern Discovery Via Toeplitz Inverse Covariance-Based Clustering

The high volatility of world soybean prices has caused uncertainty and vulnerability particularly in the developing countries. The clustering of time series is a serviceable tool for discovering soybean price patterns in temporal data. However, traditional clustering method cannot represent the continuity of price data very well, nor keep a watchful eye on the correlation between factors. In this work, the authors use the Toeplitz Inverse Covariance-Based Clustering of Multivariate Time Series Data (TICC) to soybean price pattern discovery. This is a new method for multivariate time series clustering, which can simultaneously segment and cluster the time series data. Each pattern in the TICC method is defined by a Markov random field (MRF), characterizing the interdependencies between different factors of that pattern. Based on this representation, the characteristics of each pattern and the importance of each factor can be portrayed. The work provides a new way of thinking about market price prediction for agricultural products.

SYSTEM FOR PREDICTING DISCHEARGES OVER THE HIGH WATER PERIOD THROUGH THE CLASSIFICATION TECHNIQUES DATA: CASE OF THE GAMBIA RIVER BASIN OF MAKO

This article examines the trend of flow during the high water period (from July till November) in the basin of Gambia measured at the Mako station of over 2004-2013 period. Methodology consisted at first in calculation and in standardization of data by the method of z-score of some statistical parameters (average, maximum, minimum, range and standard deviation). Obtained series were afterward submitted to classifications techniques such as k-means clustering and Agglomerative Hierarchical Clustering (AHC) of Time Series Data Mining to cluster and discover the discharge patterns in terms of the autoregressive model. From these methods, a forecast model has been developed for the discharge process on average over these years. This study presents basin flow dynamics in high water period from Time Series Data Mining technique.

An Enhanced LSTM for Trend Following of Time Series

Mining and analysis of time series data (TSD) have drawn a great concern, especially in the TSD clustering, classification, and forecast. In the industrial field, e.g., the work condition monitoring and the environmental safety, it is crucial to follow the trend of the corresponding TSD for a safety forecast, and few studies have been devoted to such a trend following. Motivated by this, we propose a trend following the strategy of TSD by using a long short-term memory (LSTM) network for safety forecast, in which the training method aggregates the particle swarm optimization (PSO) algorithm with gradient descent (GD) to obtain more competitive model parameters. Three kinds of trend representations of TSD are first defined based on the corresponding research in stock option. Then, the LSTM optimized with the PSO-GD is developed to perform the trend following. From the viewpoint of safety forecast, the trends varied in different time length are further predicted and analyzed. The superiority of the proposed algorithm is experimentally demonstrated by applying it to the electromagnetic radiation intensity TSD sampled from an actual coal mine and PM2.5 in UCI repository.

Clustering and Classification for Time Series Data in Visual Analytics: A Survey

Visual analytics for time series data has received a considerable amount of attention. Different approaches have been developed to understand the characteristics of the data and obtain meaningful statistics in order to explore the underlying processes, identify and estimate trends, make decisions and predict the future. The machine learning and visualization areas share a focus on extracting information from data. In this paper, we consider not only automatic methods but also interactive exploration. The ability to embed efficient machine learning techniques (clustering and classification) in interactive visualization systems is highly desirable in order to gain the most from both humans and computers. We present a literature review of some of the most important publications in the field and classify over 60 published papers from six different perspectives. This review intends to clarify the major concepts with which clustering or classification algorithms are used in visual analytics for time series data and provide a valuable guide for both new researchers and experts in the emerging field of integrating machine learning techniques into visual analytics.

A novel distance measure for time series: Maximum shifting correlation distance

Time series distance or similarity measurement is one of the most important problems in time series data mining, including representation, clustering, classification, and outlier detection. The existing distance measures may not efficiently deal with the drifts in the series in both the phase and amplitude. In this study, a novel measurement, maximum shifting correlation distance (MSCD), is proposed to improve the accuracy and efficiency of the time series distance measure. By integrating the curve registration and correlation, the misalignments or drifts in the phase and amplitude can be eliminated, respectively. In addition, the second distance of the MSCD (MSCD-2nd), which has the “shrinkage effect”, enhances the similarity of the samples within a cluster. The second distance has the same effect on the dynamic time warping (DTW) distance. The experimental results demonstrate that MSCD-2nd is the preferred measure in terms of the accuracy and efficiency for time series clustering and classification.

Clustering Field-Based Maize Phenotyping of Plant-Height Growth and Canopy Spectral Dynamics Using a UAV Remote-Sensing Approach.

Phenotyping under field environmental conditions is often considered as a bottleneck in crop breeding. Unmanned aerial vehicle high throughput phenotypic platform (UAV-HTPP) mounted with multi-sensors offers an efficiency, non-invasive, flexible and low-cost solution in large-scale breeding programs compared to ground investigation, especially where measurements are time-sensitive. This study was conducted at the research station of the Xiao Tangshan National Precision Agriculture Research Center of China. Using the UAV-HTPP, RGB and multispectral images were acquired during four critical growth stages of maize. We present a method of extracting plant height (PH) at the plot scale using UAV-HTPP based on the spatial structure of the maize canopy. The core steps of this method are segmentation and spatial Kriging interpolation based on multiple neighboring maximum pixels from multiple plants in a plot. Then, the relationships between the PH extracted from imagery collected using UAV-HTPP and the ground truth were examined. We developed a semi-automated pipeline for extracting, analyzing and evaluating multiple phenotypic traits: canopy cover (CC), normalized vegetation index (NDVI), PH, average growth rate of plant height (AGRPH), and contribution rate of plant height (CRPH). For these traits, we identify genotypic differences and analyze and evaluate dynamics and development trends during different maize growth stages. Furthermore, we introduce a time series data clustering analysis method into breeding programs as a tool to obtain a novel representative trait: typical curve. We classified and named nine types of typical curves of these traits based on curve morphological features. We found that typical curves can detect differences in the genetic background of traits. For the best results, the recognition rate of an NDVI typical curve is 59%, far less than the 82.3% of the CRPH typical curve. Our study provides evidence that the PH trait is among the most heritable and the NDVI trait is among the most easily affected by the external environment in maize.

Time series data intelligent clustering algorithm for landslide displacement prediction

The traditional time series data clustering for landslide displacement prediction is based on Euclidean distance measure. The time series data is clustered by distance calculation of two vectors. The correlation between components is not considered. The multiple components with single feature will interfere with the clustering results, and the accuracy of clustering results is greatly reduced. To address this problem, an intelligent clustering algorithm for time series data in landslide displacement prediction based on nonlinear dynamic time bending is proposed in this paper. By reconstructing the phase space of the landslide displacement time series, the phase space transposed matrix is obtained as the time series reconstruction matrix. After embedding dimension processing, the time series of landslide displacement is predicted by SVM data mining model. Dynamic time warping calculation is based on the correlation of time series sequence and the components. The local optimal solution is obtained by recursive search, and the whole curve path is obtained. Clustering calculation of time series data set is carried out by using hierarchical clustering algorithm according to bending path. The intelligent clustering results of time series data in landslide displacement prediction is obtained. Experimental results show that the proposed algorithm has better clustering effect and higher clustering accuracy.

Incremental fuzzy C medoids clustering of time series data using dynamic time warping distance.

Clustering time series data is of great significance since it could extract meaningful statistics and other characteristics. Especially in biomedical engineering, outstanding clustering algorithms for time series may help improve the health level of people. Considering data scale and time shifts of time series, in this paper, we introduce two incremental fuzzy clustering algorithms based on a Dynamic Time Warping (DTW) distance. For recruiting Single-Pass and Online patterns, our algorithms could handle large-scale time series data by splitting it into a set of chunks which are processed sequentially. Besides, our algorithms select DTW to measure distance of pair-wise time series and encourage higher clustering accuracy because DTW could determine an optimal match between any two time series by stretching or compressing segments of temporal data. Our new algorithms are compared to some existing prominent incremental fuzzy clustering algorithms on 12 benchmark time series datasets. The experimental results show that the proposed approaches could yield high quality clusters and were better than all the competitors in terms of clustering accuracy.

Coronal Mass Ejection Data Clustering and Visualization of Decision Trees

Abstract Coronal mass ejections (CMEs) can be categorized as either “magnetic clouds” (MCs) or non-MCs. Features such as a large magnetic field, low plasma-beta, and low proton temperature suggest that a CME event is also an MC event; however, so far there is neither a definitive method nor an automatic process to distinguish the two. Human labeling is time-consuming, and results can fluctuate owing to the imprecise definition of such events. In this study, we approach the problem of MC and non-MC distinction from a time series data analysis perspective and show how clustering can shed some light on this problem. Although many algorithms exist for traditional data clustering in the Euclidean space, they are not well suited for time series data. Problems such as inadequate distance measure, inaccurate cluster center description, and lack of intuitive cluster representations need to be addressed for effective time series clustering. Our data analysis in this work is twofold: clustering and visualization. For clustering we compared the results from the popular hierarchical agglomerative clustering technique to a distance density clustering heuristic we developed previously for time series data clustering. In both cases, dynamic time warping will be used for similarity measure. For classification as well as visualization, we use decision trees to aggregate single-dimensional clustering results to form a multidimensional time series decision tree, with averaged time series to present each decision. In this study, we achieved modest accuracy and, more importantly, an intuitive interpretation of how different parameters contribute to an MC event.