Libra: A Benchmark for Time Series Forecasting Methods

被引:9
|
作者
Bauer, Andre [1 ]
Zuefle, Marwin [1 ]
Eismann, Simon [1 ]
Grohmann, Johannes [1 ]
Herbst, Nikolas [1 ]
Kounev, Samuel [1 ]
机构
[1] Univ Wurzburg, Wurzburg, Germany
关键词
Time Series Forecasting; Benchmarking; Evaluation; COMPETITION; ACCURACY;
D O I
10.1145/3427921.3450241
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
In many areas of decision making, forecasting is an essential pillar. Consequently, there are many different forecasting methods. According to the "No-Free-Lunch Theorem", there is no single forecasting method that performs best for all time series. In other words, each method has its advantages and disadvantages depending on the specific use case. Therefore, the choice of the forecasting method remains a mandatory expert task. However, expert knowledge cannot be fully automated. To establish a level playing field for evaluating the performance of time series forecasting methods in a broad setting, we propose Libra, a forecasting benchmark that automatically evaluates and ranks forecasting methods based on their performance in a diverse set of evaluation scenarios. The benchmark comprises four different use cases, each covering 100 heterogeneous time series taken from different domains. The data set was assembled from publicly available time series and was designed to exhibit much higher diversity than existing forecasting competitions. Based on this benchmark, we perform a comprehensive evaluation to compare different existing time series forecasting methods.
引用
收藏
页码:189 / 200
页数:12
相关论文
共 50 条
  • [1] Advanced Time Series Forecasting Methods
    Egrioglu, Erol
    Khashei, Mehdi
    Aladag, Cagdas Hakan
    Turksen, I. Burhan
    Yolcu, Ufuk
    [J]. MATHEMATICAL PROBLEMS IN ENGINEERING, 2015, 2015
  • [2] Review of Time Series Traffic Forecasting Methods
    Wang, Linkai
    Chen, Jing
    Wang, Wei
    Song, Ruizhuo
    [J]. 2022 4TH INTERNATIONAL CONFERENCE ON CONTROL AND ROBOTICS, ICCR, 2022, : 419 - 423
  • [3] Time series forecasting methods in emergency contexts
    Hernandez, P. Villoria
    Marinas-Collado, I.
    Sipols, A. Garcia
    de Blas, C. Simon
    Sanchez, M. C. Rodriguez
    [J]. SCIENTIFIC REPORTS, 2023, 13 (01)
  • [4] Kernel methods applied to time series forecasting
    Rubio, Gines
    Pomares, Hetor
    Herrera, Luis J.
    Rojas, Ignacio
    [J]. COMPUTATIONAL AND AMBIENT INTELLIGENCE, 2007, 4507 : 782 - +
  • [5] Winning methods for forecasting tourism time series
    Baker, Lee C.
    Howard, Jeremy
    [J]. INTERNATIONAL JOURNAL OF FORECASTING, 2011, 27 (03) : 850 - 852
  • [6] Time series forecasting methods in emergency contexts
    P. Villoria Hernandez
    I. Mariñas-Collado
    A. Garcia Sipols
    C. Simon de Blas
    M. C. Rodriguez Sánchez
    [J]. Scientific Reports, 13
  • [7] On projection methods for functional time series forecasting
    Elias, Antonio
    Jimenez, Raul
    Shang, Han Lin
    [J]. JOURNAL OF MULTIVARIATE ANALYSIS, 2022, 189
  • [8] Time Series Neural Network Forecasting Methods
    WEN Xinhui
    CHEN Keizhou(The Centlal of Neural Netwolk
    [J]. Journal of Systems Science and Systems Engineering, 1996, (01) : 24 - 32
  • [9] TIME SERIES NEURAL NETWORK FORECASTING METHODS
    文新辉
    陈开周
    [J]. Journal of Electronics(China), 1995, (01) : 1 - 8
  • [10] Time to Focus: A Comprehensive Benchmark using Time Series Attribution Methods
    Mercier, Dominique
    Bhatt, Jwalin
    Dengel, Andreas
    Ahmed, Sheraz
    [J]. ICAART: PROCEEDINGS OF THE 14TH INTERNATIONAL CONFERENCE ON AGENTS AND ARTIFICIAL INTELLIGENCE - VOL 2, 2022, : 562 - 573