The Importance Of Time Series Analysis And Forecasting

Municipalities, market participants, customers and citizens or simply traders are all attempt to make a living from buying and selling various services either water services or financial services. Clearly the ability to forecast the direction of water consumptions and organization revenues movements, up or down, is vital to these individuals and entities. Further, over the last few decades researches and scientific papers have shown an interest in forecasting field and attempted to quantify and justify the wide variety of techniques used.

2.1. Time Series Analysis and Forecasting

A time series is a sequence of observations on a variable measured at successive points in time or over successive periods of time. The measurements may be taken every hour, day, week,

Seasonality - regular variations in the time series that is caused by re-occurring events, for example a spike in sales during the Christmas period (So & Chung, 2014).

3. Random component - additional fluctuations in the series that may be attributed to noise or other random events.

Also, Allan Steel(Steel, 2014) explained the three primary types of time series which are stationary, additive and multiplicative. Usually, stationary time series are repetitive, in other words showing constant auto-correlation and are considered the easiest type to model. Stationary series have constant amplitude without a trend element: stationary time series = seasonality +/or noise

The second type of time series is the additive type. In this type all three components of the series are present, trend, seasonality and noise. The distinguishing feature here is the amplitude of the seasonal component in that it is quite regular being static over time. This time series is trending upwards overall but there is a clear repetitive pattern of peaks and troughs caused by the seasonality, with the heights of the peaks all being similar. We can consider an additive time series as: additive time series = trend + seasonality +