Time Series - JuliaOli/SDN-Monitor GitHub Wiki
Time Series Analysis
The main objective of time series is to identify nonrandom patterns in the time series of a variable of interest, and the observation of this past behavior allow to make prevision about the future behaviors and guiding decision making. According to the classical model all time series are composed of four patterns:
- Tendencies (T), which is the long-term behavior of the series, which may be caused by demographic growth, or gradual change in spending habits, or any other aspect that affects the long-term variable of interest.
- Cyclical Variations or Cycles (C), fluctuations in the values of the variable lasting longer than one year, and recurring with certain periodicity, which may be the result of economic variations such as periods of growth or recession, or weather phenomena such as El Niño (repeating more than one year apart);
- Seasonal or Seasonal Variations (S), fluctuations in the values of the variable lasting less than one year, which are repeated every year, usually according to the seasons (or according to public holidays or popular holidays, or legal requirements, such as the period for filing the income tax return); if data are recorded annually there will be no influence of seasonality on series 3;
- Irregular irregularities (I), which are unexplained fluctuations, the result of unforeseen events such as natural disasters, terrorist attacks such as September 11, 2001 governments, etc.