A methodology to filter time series: application to minute-by-minute electric load series

Pesquisa Operacional

Mayte Suarez-farinas¹  Rodrigo Lage De Sousa¹  Reinaldo Castro Souza¹ 

[1] ,Pontifícia Univ. Católica do Rio de Janeiro Electric Engineering Department Rio de Janeiro RJ

关键词: load series;    DLM model;    irregularities in time series;    cubic spline;    Bayes factor;    série de cargas;    modelo linear dinâmico;    irregularidades em séries temporais;    spline cúbico;    fator de Bayes;   

DOI  :  10.1590/S0101-74382004000300002

来源: SciELO

【 摘 要 】

In this article a methodology for filtering a time series is presented, with application to high frequency series such as the minute-by-minute electric load series. The goal of this approach is to detect and substitute the irregularities of the time series that can produce distortions on the modelling stage. Outlier values are detected through a dynamic linear model and the Bayes factor tool; missing values are then interpolated with a Smoothing Cubic Spline. The performance of the proposed approach is illustrated using real data and evaluated through a series of tests where the irregularities have been simulated.

【 授权许可】

CC BY   
 All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License