TIME SERIES MODELLING

   

    Time series model is an econometrics model that involves time series data or variable. To understand this concept there is a need to know what time series data are? Time series data are data or observations on several variables or same variable over a given period of time or over time. However, time series is seen as  vital variables that are most common in the real life phenomenon. Such as Gross domestic product, money supply, Income of households, annual rainfall, etc or what a view.
     Thus, time series model are built so as to solve the problem attached with time series data. for instance, we might want to look at our present consumption level  which would have certainly varied or changed over time because certain factors tend to reduce or increase our consumption behaviour overtime some of these factors that may lead to changes in consumption pattern of households may be either increase or decrease in their disposable income equals PERSONAL INCOME - PERSONAL TAXATION. Or the value of Real Gross Domestic Product may increase  due to increase in inflation rate.

     Looking at time series model the use of Ordinary Least Square Method to estimating Regression  Time series Model
Say Y= Bo + B1G + U (Simple regression)
where Y = GDP
G = money supply

     Estimating this model may give a spurious regression result meaning that if it gives a good fit or tends to be significant might be misleading and giving wrong interpretation or forecasting may be detrimental.

Thus, the assumption of OLS that the mean and variance of the error term are constant does not hold true because they the variables tend to change with time which is in fact called non- stationarity and this has led to the development of time series modelling.

Hence, time series modelling is built to  solve non- stationarity of series or time variables.

Types of Time series modelling are:

1. Purely Random Process

This process is called purely random procedure because it is a discrete process consisting of a sequence of mutually independent and identically distributed random variables with the mean zero and constant variance. This process is pure because its probability distribution is not bias. It is referred to as White noise and the stochastic process is Stationary.

2. Random Walk

It involves moving from the past to the present. Its stochastic process is non-stationary.

3. Moving Average

It involves the weighted sum of the lag and current random disturbances.
TIME SERIES MODELLING TIME SERIES MODELLING Reviewed by Kolawole Victor on September 04, 2017 Rating: 5

No comments

Please don't leave without posting your comments. Its very easy folks.