GRATIS: GeneRAting TIme Series with diverse and controllable characteristics (2019-03-07T00:00:00.000000Z)

TL;DR

This work proposes GeneRAting TIme Series with diverse and controllable characteristics, named GRATIS, with the use of mixture autoregressive (MAR) models, and demonstrates the usefulness of the time series generation process through a time series forecasting application.

Abstract

The explosion of time series data in recent years has brought a flourish of new time series analysis methods, for forecasting, clustering, classification and other tasks. The evaluation of these new methods requires either collecting or simulating a diverse set of time series benchmarking data to enable reliable comparisons against alternative approaches. We propose GeneRAting TIme Series with diverse and controllable characteristics, named GRATIS, with the use of mixture autoregressive (MAR) models. We simulate sets of time series using MAR models and investigate the diversity and coverage of the generated time series in a time series feature space. By tuning the parameters of the MAR models, GRATIS is also able to efficiently generate new time series with controllable features. In general, as a costless surrogate to the traditional data collection approach, GRATIS can be used as an evaluation tool for tasks such as time series forecasting and classification. We illustrate the usefulness of our time series generation process through a time series forecasting application.

Authors

Yanfei Kang

1 papers

Rob J Hyndman

1 papers

Feng Li

1 papers

TL;DR

Abstract

Authors

References78 items

Tree-based methods for clustering time series using domain-relevant attributes

The M4 Competition: Results, findings, conclusion and way forward

Deep learning for time series classification: a review

Instance spaces for machine learning classification

Mastering the game of Go without human knowledge

Feature-based time-series analysis

Evaluation and comparison of statistical methods for early temporal detection of outbreaks: A simulation-based study

Generating What-If Scenarios for Time Series Data

Visualising forecasting algorithm performance using time series instance spaces

Simulated Data Experiments for Time Series Classification Part 1: Accuracy Comparison with Default Settings

PMLB: a large benchmark suite for machine learning evaluation and comparison

Performance Analysis of Continuous Black-Box Optimization Algorithms via Footprints in Instance Space

Large-Scale Unusual Time Series Detection

Generating new test instances by evolving in instance space

An Iterative Coordinate Descent Algorithm for High-Dimensional Nonconvex Penalized Quantile Regression

Classes of structures in the stable atmospheric boundary layer

'Horses for Courses' in demand forecasting

Detecting and Classifying Events in Noisy Time Series

psd: Adaptive, sine multitaper power spectral density estimation for R

Highly Comparative Feature-Based Time-Series Classification

Forecasting: principles and practice

Horses for courses

Approximation of conditional densities by smooth mixtures of regressions

Flexible Modeling of Conditional Distributions Using Smooth Mixtures of Asymmetric Student T Densities

How does improved forecasting benefit detection? An application to biosurveillance

ImageNet: A large-scale hierarchical image database

Rule induction for forecasting method selection: Meta-learning the characteristics of univariate time series

Seasonality and the Order of Integration for Consumption

Maximum Entropy Bootstrap for Time Series: The meboot R Package

Et al

Explaining and Forecasting Online Auction Prices and Their Dynamics Using Functional Data Analysis

Regression Density Estimation Using Smooth Adaptive Gaussian Mixtures

Characteristic-Based Clustering for Time Series Data

Another look at measures of forecast accuracy

Bayesian feature and model selection for Gaussian mixture models

Analysis of low count time series data by poisson autoregression

Time series classification using Gaussian mixture models of reconstructed phase spaces

Hierarchical Topic Models and the Nested Chinese Restaurant Process

On the Need for Time Series Data Mining Benchmarks: A Survey and Empirical Demonstration

A simulation study of artificial neural networks for nonlinear time-series forecasting

Evidence for the selection of forecasting methods

On the Approximation Rate of Hierarchical Mixtures-of-Experts for Generalized Linear Models

Existence and Stochastic Structure of a Non‐negative Integer‐valued Autoregressive Process

Model selection in univariate time series forecasting using discriminant analysis

No free lunch theorems for optimization

Modeling flat stretches, bursts, and outliers in time series using mixture transition distribution models

The Lack of A Priori Distinctions Between Learning Algorithms

Bayesian Density Estimation and Inference Using Mixtures

“Rule-based forecasting: Development and validation of an expert systems approach to combining time series extrapolations”: Fred Collopy and J. Scott Armstrong, Management science, 38 (1992) 1394–1414

Testing the null hypothesis of stationarity against the alternative of a unit root: How sure are we that economic time series have a unit root?

Rule-Based Forecasting: Development and Validation of an Expert Systems Approach to Combining Time Series Extrapolations

Multivariate adaptive regression splines

Space-time modeling with long-memory dependence: assessing Ireland's wind-power resource. Technical report

A VARIABLE SPAN SMOOTHER

Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation

On a measure of lack of fit in time series models

INDIVIDUAL ITEM FORECASTING MODEL EVALUATION

Are forecasting competitions data representative of the reality?

Space-time Modelling with Long-memory Dependence: Assessing Ireland's Wind Power Resource

R Core Team

shiny: Web Application Framework for R

tsfeatures: Time Series Feature Extraction . Version 0.1

Instance spaces for machine learning classification

UCI Machine Learning Repository

and E

Tcomp: Data from the 2010 Tourism Forecasting Competition

Atlantic City, NJ, USA

R: A language and environment for statistical computing.

Bayesian data analysis, third edition

Approximation of conditional densities by smooth mixtures of regressions ∗

Simulating and Evaluating Biosurveillance Datasets

Visualizing Data using t-SNE

Regression Quantiles

Statistical Society: Series B (Statistical Methodology

On a mixture autoregressive model

STL : A Seasonal-Trend Decomposition Procedure Based on Loess