Clustering Of Time Series Data Research Articles

Time series clustering of Malaysia Air quality time series data

Air quality is often associated with the area location and activities where air quality in cities is usually more polluted than in rural areas. This study aims to study the pattern of time series data from air quality stations by performing cluster analysis of air quality station based on the particulate matter 10 micrometres or less in diameter (PM10) and particulate matter 2.5 micrometres or less in diameter (PM2.5) time series data. The clusters obtained from the cluster analysis were compared with the station area category and station location. This study which uses air quality data obtained from the Department of Environment, Malaysia from 5 July 2017 until 30 June 2019, shows five types of air quality patterns in Malaysia. The results also show that none of the clusters is dominated by any station's category. Therefore, it is less appropriate to relate the air quality patterns and the station area category. However, the results show that air quality patterns were related to the station's location, where nearby stations have similar air quality patterns.

A modified Kohonen map algorithm for clustering time series data

Time Series clustering is a domain with several applications spanning various fields. The concept of vector quantization, popularly used in signal processing to approximate a large number of signals, can be used to cluster signals and thereby time series data. Though a popular clustering algorithm such as K-Means is capable of performing vector quantization, the averaging technique to compute centroids in the algorithm is not well suited to handle time series data. The ability of Self Organizing Map algorithm, has, therefore, been explored in this work to perform clustering of time series data by adopting several modifications in the original steps of the algorithm. By initializing the prototype vectors using a farthest neighbors’ approach instead of random initialization and using the dynamic time warping distance measure to calculate similarity between signals, a novel procedure has been proposed to apply the Self Organizing Map algorithm to cluster time series data. The proposed algorithm is first tested on 119 data sets and its performance is compared to that of Agglomerative Clustering and k medoids clustering using 3 validation measures. Next, their scalability is compared by looking at their time of computation on the data sets. Performance of the proposed algorithm in terms of the fluctuations involved due to initialization and the parameters of the algorithm are studied next using 3 more validation measures. The results showcase that the modified Self Organizing Map is not only a better algorithm than Agglomerative Clustering in terms of clustering performance, but also more scalable in terms of taking less time to compute clusters as it performs them in lesser time that k medoids while having similar cluster quality.

Optimized time series data clustering method for NN network

Time series data is a common sort of data that can be found in a variety of fields. Clustering time series data has a wide range of uses and has drawn scholars from several fields. A unique approach for shape-based time series clustering is proposed in this research. Using the complex network principle, it may reduce the bulk of data, enhance efficiency, and not reduce the effects. To begin, a one-nearest neighbour network is constructed using time series items that are comparable. The similarity is measured in this step using the triangle distance. Each node in the neighbour network represents a single time series object, whereas each connection indicates a neighbour relationship between nodes. Second, high-degree nodes are selected and used to cluster. Dynamic time warping distance function and hierarchical clustering algorithm are used in the clustering process. Finally, several studies are carried out using both synthetic and actual data. The findings demonstrate that the suggested approach is both efficient and effective.

A novel time series data clustering approach for wind speed forecasting

Wind energy plays an essential role in the generation process of sustainable energy, with a bright future. Therefore, predicting wind speed fluctuations and their output power plays a crucial role in electric power generation. The integration of wind power is based on the accuracy of wind speed and power prediction model. In this paper, a clustering algorithm is proposed based on the length of the trendlet components. After spotting the different clusters, one suitable cluster is selected for modeling using the panda’s correlation method. This paper uses specific ARIMA, Naive Forecast, and Holt Winter models to forecast the selected cluster. Here three hybrid models, namely, C-ARIMA, C-NAIVE Forecast, and C-Holt-Winter, are proposed for wind speed forecasting. The performances of the proposed models are evaluated using the mean absolute error (MAE) and root mean squared error (RMSE). The experiment outcomes show that the cluster-based forecasting technique (Hybrid models) improved performance compared with un-clustered forecasting techniques.

Behavior Algorithm — a Novel Time-series Clustering Approach

Clustering time-series values is an established technique for organizations employing machine learning to analyze temporal datasets. Generally speaking, the goal of time-series methodology is to generate predictions. Such predictions could help organizations understand potential future cyberattacks, financial market changes, weather, or disease outbreaks. However, computational limitations lead existing algorithms to fail to group individual series together based the actual behavior of the series. A feature that can be used or derived to explain the time-series behavior had not been identified in the literature despite there being a need to have numeric values to describe the pattern of values over time. To address this gap, this work presents a behavior algorithm which addresses clustering time-series data based solely on the behavior of the series. Further, the algorithm is designed to operate effectively regardless of absolute values or temporal shifts. First, we describe the algorithm through mathematical examples. We provide the design approach for the algorithm numerically and through data visualizations. Then, we validated the algorithm on sample random data. Finally, we offer conclusions along with notions for future work based on this study.

Kernel Distance Measures for Time Series, Random Fields and Other Structured Data

This paper introduces kdiff, a novel kernel-based measure for estimating distances between instances of time series, random fields and other forms of structured data. This measure is based on the idea of matching distributions that only overlap over a portion of their region of support. Our proposed measure is inspired by MPdist which has been previously proposed for such datasets and is constructed using Euclidean metrics, whereas kdiff is constructed using non-linear kernel distances. Also, kdiff accounts for both self and cross similarities across the instances and is defined using a lower quantile of the distance distribution. Comparing the cross similarity to self similarity allows for measures of similarity that are more robust to noise and partial occlusions of the relevant signals. Our proposed measure kdiff is a more general form of the well known kernel-based Maximum Mean Discrepancy distance estimated over the embeddings. Some theoretical results are provided for separability conditions using kdiff as a distance measure for clustering and classification problems where the embedding distributions can be modeled as two component mixtures. Applications are demonstrated for clustering of synthetic and real-life time series and image data, and the performance of kdiff is compared to competing distance measures for clustering.

Deep Time-Series Clustering: A Review

We present a comprehensive, detailed review of time-series data analysis, with emphasis on deep time-series clustering (DTSC), and a case study in the context of movement behavior clustering utilizing the deep clustering method. Specifically, we modified the DCAE architectures to suit time-series data at the time of our prior deep clustering work. Lately, several works have been carried out on deep clustering of time-series data. We also review these works and identify state-of-the-art, as well as present an outlook on this important field of DTSC from five important perspectives.

Agglomerative likelihood clustering

We consider the problem of fast time-series data clustering. Building on previous work modeling, the correlation-based Hamiltonian of spin variables we present an updated fast non-expensive agglomerative likelihood clustering algorithm (ALC). The method replaces the optimized genetic algorithm based approach (f-SPC) with an agglomerative recursive merging framework inspired by previous work in econophysics and community detection. The method is tested on noisy synthetic correlated time-series datasets with a built-in cluster structure to demonstrate that the algorithm produces meaningful non-trivial results. We apply it to time-series datasets as large as 20 000 assets and we argue that ALC can reduce computation time costs and resource usage costs for large scale clustering for time-series applications while being serialized, and hence has no obvious parallelization requirement. The algorithm can be an effective choice for state-detection for online learning in a fast non-linear data environment, because the algorithm requires no prior information about the number of clusters.

A Hybrid Approach for Clustering Uncertain Time Series

Information uncertainty extensively exists in the real-world applications, and uncertain data process and analysis have been a crucial issue in the area of data and knowledge engineering. In this paper, we concentrate on uncertain time series data clustering, in which the uncertain values at time points are represented by probability density function. We propose a hybrid clustering approach for uncertain time series. Our clustering approach first partitions the uncertain time series data into a set of micro-clusters and then merges the micro-clusters following the idea of hierarchical clustering. We evaluate our approach with experiments. The experimental results show that, compared with the traditional UK-means clustering algorithm, the Adjusted Rand Index (ARI) of our clustering results have an obviously higher accuracy. In addition, the time efficiency of our clustering approach is significantly improved.

End-to-end deep representation learning for time series clustering: a comparative study

Time series are ubiquitous in data mining applications. Similar to other types of data, annotations can be challenging to acquire, thus preventing from training time series classification models. In this context, clustering methods can be an appropriate alternative as they create homogeneous groups allowing a better analysis of the data structure. Time series clustering has been investigated for many years and multiple approaches have already been proposed. Following the advent of deep learning in computer vision, researchers recently started to study the use of deep clustering to cluster time series data. The existing approaches mostly rely on representation learning (imported from computer vision), which consists of learning a representation of the data and performing the clustering task using this new representation. The goal of this paper is to provide a careful study and an experimental comparison of the existing literature on time series representation learning for deep clustering. In this paper, we went beyond the sole comparison of existing approaches and proposed to decompose deep clustering methods into three main components: (1) network architecture, (2) pretext loss, and (3) clustering loss. We evaluated all combinations of these components (totaling 300 different models) with the objective to study their relative influence on the clustering performance. We also experimentally compared the most efficient combinations we identified with existing non-deep clustering methods. Experiments were performed using the largest repository of time series datasets (the UCR/UEA archive) composed of 128 univariate and 30 multivariate datasets. Finally, we proposed an extension of the class activation maps method to the unsupervised case which allows to identify patterns providing highlights on how the network clustered the time series.

Building Energy Time Series Data Mining for Behavior Analytics and Forecasting Energy consumption

The significant aim of this research has always been to evaluate the mechanism for efficient and inherently aware usage of vitality in-home devices, thus improving the information of smart metering systems with regard to the usage of selected homes and the time of use. Advances in information processing are commonly used to quantify gigantic building activity data steps to boost the activity efficiency of the building energy systems. Here, some smart data mining models are offered to measure, and predict the time series for energy in order to expose different ephemeral principles for using energy. Such considerations illustrate the use of machines in relation to time, such as day hour, time of day, week, month and year relationships within a family unit, which are key components in gathering and separating the effect of consumers behaviors in the use of energy and their pattern of energy prediction. It is necessary to determine the multiple relations through the usage of different appliances from simultaneous information flows. In comparison, specific relations among interval-based instances where multiple appliances use continue for certain duration are difficult to determine. In order to resolve these difficulties, an unsupervised energy time-series data clustering and a frequent pattern mining study as well as a deep learning technique for estimating energy use were presented. A broad test using true data sets that are rich in smart meter data were conducted. The exact results of the appliance designs that were recognized by the proposed model were filled out by Deep Convolutional Neural Networks (CNN) and Recurrent Neural Networks (LSTM and GRU) at each stage, with consolidated accuracy of 94.79%, 97.99%, 99.61%, for 25%, 50%, and 75%, respectively.

A Bayesian method for on-line evaluation of uncertainty in measurement of Coriolis flow meters

On-line evaluation of measurement uncertainty is crucial for process control and quality control in real applications. Traditional approaches to measurement uncertainty (MU) assume that measurands are repeated measurements collected in static laboratory conditions. On-line evaluation of MU, then, constitutes a challenging problem because the sensor data is collected under a variety of operating conditions. We propose a new method for the on-line evaluation of MU which consists of clustering time series data into groups with similar operational conditions and evaluating the MU using Bayesian inference. The mass count uncertainty measured using Coriolis mass flow meters on two hydrogen refueling stations is evaluated. The clustering of fueling events effectively reduces the process noise in on-line evaluation, and the Bayesian inference method identifies a much narrower uncertainty range than conventional methods. Therefore, our approach of using machine learning methods for on-line evaluation of MU is a promising practical approach.

Fault diagnosis based on feature clustering of time series data for loss and kick of drilling process

With the increase of drilling depth, complicated geological environments lead to a high risk of loss and kick. Fault diagnosis plays an essential role in minimizing the financial and environmental losses of the drilling process. On account of the temporal correlation of drilling parameters, a fault diagnosis method based on feature clustering of time series data for loss and kick of the drilling process is presented in this paper. Distance correlation is conducted for parameter combination to retain the whole information of drilling process. Global trend, local trends, and approximate entropy features are extracted to illustrate the characteristic of the time series. Density-based clustering method is performed for each combination to mine the local similarity among drilling parameters. Based on the clustering results of each combination as the inputs, the Bayesian classifier is further utilized to obtain the final fault diagnosis result. Experiments are executed with the actual data collected from a practical drilling process. The results indicate that the proposed method has both low false alarm rate and low miss alarm rate.

Clustering of Commodity Inflation Pattern based on Estimated ARIMA Model

Clustering is the stage that is carried out before further data analysis. There are many approach method that can be used for clustering time series data, one of which is the model-based approach. In this study, we clustering inflation data used the ARIMA model. The cluster model is carried out after obtaining the clusters. The similarity between time series is measured using the development of Piccolo distance. Furthermore, the Ward hierarchical method is used for clustering. The Silhouette averaging method is used to determine the optimal number of clusters. The cluster model can be used to represent individual models. The cluster model is more effective than creating all individual models.

Predicting and forecasting of time series models using cluster analysis

Time series model on multiple objects could be univariate and multivariate model. The more objects used in multivariate model would decrease precision of forecast for each object. One way to overcome this problem used univariate models for each object. However, univariate models for each object became inefficient in time. Therefore, clustering performed on objects so that model became efficient. The objective of this research is to study results of predicting and forecasting model with and without clustering. Model and forecasting used time series regression model on broad proportion of plant-disturbing organism attack and planting area of food crops in Indonesia. Clustering of time series data was same as clustering in general, but the distance and method should be able to accommodate time series data structure which was dynamic in time. Evaluation of prediction and forecast mean average percentage error (MAPE) show that forecast of model was performed with clustering as good as forecast of model for each object. However, prediction of model with clustering was not as good as the prediction of model for each object so that the prediction of broad proportion of plant-disturbing organism attack only served as an indicator of the arrival populations of plant-disturbing organism.

Morphological Similarity Clustering and Its Applications in Anomaly Detection of Time Series

Time series data clustering is an important branch and difficult topic in the field of data clustering. In this paper, the definition of temporal data morphological similarity is proposed, a set of affine invariant morphological similarity measurement methods of time series data is established, and a morphological clustering algorithm based on morphological similarity measurement is developed. Using morphological similarity measurement of time series data, two groups of abnormal change detection algorithms for time series data are established, which can be used to detect the morphological consistency of different periodical sampling series in the same time series and the morphological consistency among several time series in the same period. Based on these algorithms stated above, the multiple monitoring algorithms are proposed, which can be used to monitor states of many kinds of industry process. The effectiveness of the methods and algorithms is verified with theoretical deduction and simulation results. Simulation results show that these algorithms are very valuable for mining, clustering, modeling, statistical learning of multi-source time series data, as well as the detection and diagnosis of abnormal process changes.

Multivariate Time Series Data Clustering Method Based on Dynamic Time Warping and Affinity Propagation

In view of the importance of various components and asynchronous shapes of multivariate time series, a clustering method based on dynamic time warping and affinity propagation is proposed. From the two perspectives of the global and local properties information of multivariate time series, the relationship between the data objects is described. It uses dynamic time warping to measure the similarity between original time series data and obtain the similarity between the corresponding components. Moreover, it also uses the affinity propagation to cluster based on the similarity matrices and, respectively, establishes the correlation matrices for various components and the whole information of multivariate time series. In addition, we further put forward the synthetical correlation matrix to better reflect the relationship between multivariate time series data. Again the affinity propagation algorithm is applied to clustering the synthetical correlation matrix, which realizes the clustering analysis of the original multivariate time series data. Numerical experimental results demonstrate that the efficiency of the proposed method is superior to the traditional ones.

What is the busiest time at an airport? Clustering U.S. hub airports based on passenger movements

Accurate clustering of airports enables airport authorities and operators to precisely position themselves in the competitive market, facilitates airlines to efficiently allocate resources, and empowers the research community with credible data for new discoveries. Existing clustering studies largely use traffic volume, network connectivity, or operational efficiency to group airports into hierarchical clusters or parallel partitions. Developing a different perspective, this study focuses upon a key bottleneck of outbound passenger movements in airport terminals, the security checkpoints, and uses the Transportation Security Administration (TSA) Customer Throughput/Wait Times Reports to cluster hub airports of the United States. With the input data structured as univariate time-series, k-shape, a clustering algorithm that is robust to time axis distortion and computationally efficient, is selected to analyze the similarity of time-series using shape-based distance. The clustering results are validated by examining the raw and z-normalized data of selected airport clusters on six sampled dates. Analysis results indicate that k-shape is competent and efficient to process and cluster time-series data used for this specific research. This study offers a fresh perspective to cluster commercial airports using an infrequently employed dataset. The clustering results reveal how the geographical location, hub status in airlines' operational network, and destination type of an airport affect the movement of outbound passengers through terminals

Time Series Clustering Based on the K-Means Algorithm

Time series is one of the forms of data presentation that is used in many studies. It is convenient, easy and informative. Clustering is one of the tasks of data processing. Thus, the most relevant currently are methods for clustering time series. Clustering time series data aims to create clusters with high similarity within a cluster and low similarity between clusters. This work is devoted to clustering time series. Various methods of time series clustering are considered. Examples are given for real data.

Time-series Data Clustering Based on the Correlation of Periodogram

Time-series Data Clustering Based on the Correlation of Periodogram