Multiple Time Series

14,000,000 Leading Edge Experts on the ideXlab platform

Scan Science and Technology

Contact Leading Edge Experts & Companies

Scan Science and Technology

Contact Leading Edge Experts & Companies

The Experts below are selected from a list of 273 Experts worldwide ranked by ideXlab platform

Costas J. Spanos - One of the best experts on this subject based on the ideXlab platform.

  • CDC - BRIEF: Bayesian Regression of Infinite Expert Forecasters for single and Multiple Time Series prediction
    2015 54th IEEE Conference on Decision and Control (CDC), 2015
    Co-Authors: Costas J. Spanos
    Abstract:

    Bayesian Regression of Infinite Expert Forecasters (BRIEF) as proposed in the study is a prediction algorithm for Time-varying systems. The method is based on regret minimization by tracking the performance of an inifinite pool of experts for single and Multiple Time Series. The inverse correlation weighted error (ICWE) employed in BRIEF takes into account the dependency structure among Multiple Time Series, which can also be adapted to multi-step ahead predictions. Theoretical bounds show that the cumulative regret grows at rate O(log T) with respect to the oracle that can select the best strategy in retrospect. As the per round regret vanishes, BRIEF is indistinguishable to the oracle when the horizon increases. Also since the bound applies to any choice of input subject to the euclidean norm constraint, the method can be applied to adversarial settings. Experimental results verify that BRIEF excels in single and Multiple steps ahead prediction of ARMAX simulated data and building energy consumptions.

  • BRIEF: Bayesian Regression of Infinite Expert Forecasters for single and Multiple Time Series prediction
    2015 54th IEEE Conference on Decision and Control (CDC), 2015
    Co-Authors: Costas J. Spanos
    Abstract:

    Bayesian Regression of Infinite Expert Forecasters (BRIEF) as proposed in the study is a prediction algorithm for Time-varying systems. The method is based on regret minimization by tracking the performance of an inifinite pool of experts for single and Multiple Time Series. The inverse correlation weighted error (ICWE) employed in BRIEF takes into account the dependency structure among Multiple Time Series, which can also be adapted to multi-step ahead predictions. Theoretical bounds show that the cumulative regret grows at rate O(log T) with respect to the oracle that can select the best strategy in retrospect. As the per round regret vanishes, BRIEF is indistinguishable to the oracle when the horizon increases. Also since the bound applies to any choice of input subject to the euclidean norm constraint, the method can be applied to adversarial settings. Experimental results verify that BRIEF excels in single and Multiple steps ahead prediction of ARMAX simulated data and building energy consumptions.

Nikola Kasabov - One of the best experts on this subject based on the ideXlab platform.

  • Dynamic Learning of Multiple Time Series in a Nonstationary Environment
    Learning in Non-Stationary Environments, 2020
    Co-Authors: Harya Widiputra, Russel Pears, Nikola Kasabov
    Abstract:

    This chapter introduces two distinct solutions to the problem of capturing the dynamics of Multiple Time Series and the extraction of useful knowledge over Time. As these dynamics would change in a nonstationary environment, the key characteristic of the methods is the ability to evolve their structure continuously over Time. In addition, reviews of existing methods of dynamic single Time Series analysis and modeling such as the dynamic neuro-fuzzy inference system and the neuro-fuzzy inference method for transductive reasoning, which inspired the proposed methods, are presented. This chapter also presents a comprehensive evaluation of the performance of the proposed methods on a real-world problem, which consists of predicting movement of global stock market indexes over Time.

  • Evolving integrated multi-model framework for on line Multiple Time Series prediction
    Evolving Systems, 2013
    Co-Authors: Russel Pears, Harya Widiputra, Nikola Kasabov
    Abstract:

    Time Series prediction has been extensively researched in both the statistical and computational intelligence literature with robust methods being developed that can be applied across any given application domain. A much less researched problem is Multiple Time Series prediction where the objective is to simultaneously forecast the values of Multiple variables which interact with each other in Time varying amounts continuously over Time. In this paper we describe the use of a novel integrated multi-model framework (IMMF) that combined models developed at three different levels of data granularity, namely the global, local and transductive models to perform Multiple Time Series prediction. The IMMF is implemented by training a neural network to assign relative weights to predictions from the models at the three different levels of data granularity. Our experimental results indicate that IMMF significantly outperforms well established methods of Time Series prediction when applied to the Multiple Time Series prediction problem.

  • PAKDD (2) - Multiple Time-Series prediction through Multiple Time-Series relationships profiling and clustered recurring trends
    Advances in Knowledge Discovery and Data Mining, 2011
    Co-Authors: Harya Widiputra, Russel Pears, Nikola Kasabov
    Abstract:

    Time-Series prediction has been very well researched by both the Statistical and Data Mining communities. However the Multiple Time-Series problem of predicting simultaneous movement of a collection of Time sensitive variables which are related to each other has received much less attention. Strong relationships between variables suggests that trajectories of given variables that are involved in the relationships can be improved by including the nature and strength of these relationships into a prediction model. The key challenge is to capture the dynamics of the relationships to reflect changes that take place continuously over Time. In this research we propose a novel algorithm for extracting profiles of relationships through an evolving clustering method. We use a form of non-parametric regression analysis to generate predictions based on the profiles extracted and historical information from the past. Experimental results on a real-world climatic data reveal that the proposed algorithm outperforms well established methods of Time-Series prediction.

  • DYNAMIC INTERACTION NETWORKS VERSUS LOCAL TREND MODELS FOR Multiple Time-Series PREDICTION
    Cybernetics and Systems, 2011
    Co-Authors: Harya Widiputra, Russel Pears, Nikola Kasabov
    Abstract:

    Time-Series modeling and prediction have been very well researched by both the statistical and data mining communities. However, the Multiple Time-Series problem of modeling and predicting simultaneous movements of a collection of Time-sensitive variables that are related to each other has received much less attention. Strong relationships between variables suggest that trajectories of given variables involved in the relationships can be improved by including the nature and strength of these relationships in a prediction model. The key challenge is to capture the dynamics of the relationships to reflect changes that take place continuously over Time. This research presents a global model to capture inclusive patterns of dynamic interactions between Multiple Time-Series and a local trend model to extract localized profiles of relationships and recurring trends in Multiple Time-Series. Our experimentation revealed that the global and local models specially developed for Multiple Time-Series prediction outperformed methods such as Multiple linear regression and the multilayer perceptron that were developed for predicting single Time-Series.

Xuerui Wang - One of the best experts on this subject based on the ideXlab platform.

  • dynamic mixture models for Multiple Time Series
    International Joint Conference on Artificial Intelligence, 2007
    Co-Authors: Xuerui Wang
    Abstract:

    Traditional probabilistic mixture models such as Latent Dirichlet Allocation imply that data records (such as documents) are fully exchangeable. However, data are naturally collected along Time, thus obey some order in Time. In this paper, we present Dynamic Mixture Models (DMMs) for online pattern discovery in Multiple Time Series. DMMs do not have the noticeable drawback of the SVD-based methods for data streams: negative values in hidden variables are often produced even with all non-negative inputs. We apply DMM models to two real-world datasets, and achieve significantly better results with intuitive interpretation.

  • IJCAI - Dynamic mixture models for Multiple Time Series
    2007
    Co-Authors: Xuerui Wang
    Abstract:

    Traditional probabilistic mixture models such as Latent Dirichlet Allocation imply that data records (such as documents) are fully exchangeable. However, data are naturally collected along Time, thus obey some order in Time. In this paper, we present Dynamic Mixture Models (DMMs) for online pattern discovery in Multiple Time Series. DMMs do not have the noticeable drawback of the SVD-based methods for data streams: negative values in hidden variables are often produced even with all non-negative inputs. We apply DMM models to two real-world datasets, and achieve significantly better results with intuitive interpretation.

Yun-feng Zhao - One of the best experts on this subject based on the ideXlab platform.

  • A New Model for Multiple Time Series Based on Data Mining
    2008 International Symposium on Knowledge Acquisition and Modeling, 2008
    Co-Authors: Zhuo Chen, Bing-ru Yang, Fa-guo Zhou, Lin-na Li, Yun-feng Zhao
    Abstract:

    Time Series are the important type of data in the world, and Time Series data mining is one of the most important subfields of data mining. In this paper we propose a model of temporal pattern discovery from Multiple Time Series based on temporal logic. Firstly, Multiple Time Series are transform to Multiple event sequences, and then they are synthesized into one event sequence. Secondly, we generate the observation sequence to mining the temporal pattern and the rules based on the interval temporal logic. The algorithm is proposed to mining online frequent episodes and mining change of patterns on mass event sequences. Finally, efficiency of the model and the algorithm is proved through experiments.

Harya Widiputra - One of the best experts on this subject based on the ideXlab platform.

  • Dynamic Learning of Multiple Time Series in a Nonstationary Environment
    Learning in Non-Stationary Environments, 2020
    Co-Authors: Harya Widiputra, Russel Pears, Nikola Kasabov
    Abstract:

    This chapter introduces two distinct solutions to the problem of capturing the dynamics of Multiple Time Series and the extraction of useful knowledge over Time. As these dynamics would change in a nonstationary environment, the key characteristic of the methods is the ability to evolve their structure continuously over Time. In addition, reviews of existing methods of dynamic single Time Series analysis and modeling such as the dynamic neuro-fuzzy inference system and the neuro-fuzzy inference method for transductive reasoning, which inspired the proposed methods, are presented. This chapter also presents a comprehensive evaluation of the performance of the proposed methods on a real-world problem, which consists of predicting movement of global stock market indexes over Time.

  • Nearest neighbour approach with non-parametric regression analysis for Multiple Time-Series modelling and predictions
    International Journal of Business Intelligence and Data Mining, 2015
    Co-Authors: Harya Widiputra
    Abstract:

    Time-Series prediction is an intensively researched area, yet most studies in this field have focused on predicting movements of a single Series only, whilst prediction of Multiple Time-Series based on patterns of interaction between Multiple Time-Series has received very little attention. On the other hand, findings in various studies show that given a Multiple Time-Series data there exist patterns of relationship between the observed variables, and being able to model them would lead to the possibility of building a more accurate model to predict their future values. Nevertheless, as real-world systems change dynamically over Time, having a single model to explain simultaneous movement of Multiple Time-Series will not be sufficient. To address this problem, the paper presents an algorithm that is capable of building a new decision model on-the-fly based on the state of relationships, between observed variables at a particular Time-point. The proposed algorithm utilises non-parametric regression analysis to extract profiles of relationship between observed variables and then employs the nearest neighbour approach to find appropriate conditions from the past. Experimental results on a real-world dataset suggest that the implementation of kernel regression merged with nearest neighbour approach shows that it outperforms established methods such as Multiple linear regression and multi-layer perceptron.

  • Evolving integrated multi-model framework for on line Multiple Time Series prediction
    Evolving Systems, 2013
    Co-Authors: Russel Pears, Harya Widiputra, Nikola Kasabov
    Abstract:

    Time Series prediction has been extensively researched in both the statistical and computational intelligence literature with robust methods being developed that can be applied across any given application domain. A much less researched problem is Multiple Time Series prediction where the objective is to simultaneously forecast the values of Multiple variables which interact with each other in Time varying amounts continuously over Time. In this paper we describe the use of a novel integrated multi-model framework (IMMF) that combined models developed at three different levels of data granularity, namely the global, local and transductive models to perform Multiple Time Series prediction. The IMMF is implemented by training a neural network to assign relative weights to predictions from the models at the three different levels of data granularity. Our experimental results indicate that IMMF significantly outperforms well established methods of Time Series prediction when applied to the Multiple Time Series prediction problem.

  • PAKDD (2) - Multiple Time-Series prediction through Multiple Time-Series relationships profiling and clustered recurring trends
    Advances in Knowledge Discovery and Data Mining, 2011
    Co-Authors: Harya Widiputra, Russel Pears, Nikola Kasabov
    Abstract:

    Time-Series prediction has been very well researched by both the Statistical and Data Mining communities. However the Multiple Time-Series problem of predicting simultaneous movement of a collection of Time sensitive variables which are related to each other has received much less attention. Strong relationships between variables suggests that trajectories of given variables that are involved in the relationships can be improved by including the nature and strength of these relationships into a prediction model. The key challenge is to capture the dynamics of the relationships to reflect changes that take place continuously over Time. In this research we propose a novel algorithm for extracting profiles of relationships through an evolving clustering method. We use a form of non-parametric regression analysis to generate predictions based on the profiles extracted and historical information from the past. Experimental results on a real-world climatic data reveal that the proposed algorithm outperforms well established methods of Time-Series prediction.

  • DYNAMIC INTERACTION NETWORKS VERSUS LOCAL TREND MODELS FOR Multiple Time-Series PREDICTION
    Cybernetics and Systems, 2011
    Co-Authors: Harya Widiputra, Russel Pears, Nikola Kasabov
    Abstract:

    Time-Series modeling and prediction have been very well researched by both the statistical and data mining communities. However, the Multiple Time-Series problem of modeling and predicting simultaneous movements of a collection of Time-sensitive variables that are related to each other has received much less attention. Strong relationships between variables suggest that trajectories of given variables involved in the relationships can be improved by including the nature and strength of these relationships in a prediction model. The key challenge is to capture the dynamics of the relationships to reflect changes that take place continuously over Time. This research presents a global model to capture inclusive patterns of dynamic interactions between Multiple Time-Series and a local trend model to extract localized profiles of relationships and recurring trends in Multiple Time-Series. Our experimentation revealed that the global and local models specially developed for Multiple Time-Series prediction outperformed methods such as Multiple linear regression and the multilayer perceptron that were developed for predicting single Time-Series.