Scribd
Upload
4 views3 pages

RSP Assignment

The document discusses heteroscedasticity, which indicates non-constant variances in observations, affecting the efficiency of Ordinary Least Squares (OLS) estimators, and outlines various tests for detection and correction methods. It also explains stationarity in time series data, emphasizing the need for constant mean and variance over time, and describes methods for transforming non-stationary data into stationary data, such as differencing and logarithmic transformations. Key identification techniques for non-stationarity include the ACF plot.

Uploaded by

diabhatnagar20
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
4 views3 pages

RSP Assignment

The document discusses heteroscedasticity, which indicates non-constant variances in observations, affecting the efficiency of Ordinary Least Squares (OLS) estimators, and outlines various tests for detection and correction methods. It also explains stationarity in time series data, emphasizing the need for constant mean and variance over time, and describes methods for transforming non-stationary data into stationary data, such as differencing and logarithmic transformations. Key identification techniques for non-stationarity include the ACF plot.

Uploaded by

diabhatnagar20
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

RSP Assignment

Aakash Bhatnagar
BBA(FIA) 2A
Roll No.-17319

Q.1-Heteroscedasticity
Ans.1- Heteroscedasticity refers to the property of the observations, if do not
have constant variances and hence, their error terms could be varying by huge
margins, or if they do not exhibit randomness and hence, the error terms
show a fixed pattern of arrangement(Increasing, decreasing, linear, conic,etc.)
Homoscedasticity is one of the assumptions of Original least squares method
of Regressions analysis. It assumes that the data has error terms which are
normally distributed/random and hence, no heteroscedasticity is present.
If this assumption is violated, then the estimators calculated by the OLS
method would not be efficient and covariances would be underestimated,
although it would still remain unbiased. This makes the measurement of
population parameters difficult, due to higher variances.
This leads to biased standard error, and hence, biased inferences drawn from
the data,
There are various tests in order to detect heteroscedasticity in the data. Some
of them are-
1)Levene test
2)Goldfeld-Quandt test
3)Park test
4)Glejser test
5)Brown-Forsythe test
6)Breusch-Pagan test
7)White test
In case heteroscedasticity is detected in the data, there are various measures
in order to correct it as well-
1)we could change the functional form of the data, by taking log on both
sides of the model
2)use different specifications of the model
3)Box-Cox transformation

Q.2-Stationarity
Ans.2- In statistics, stationarity refers to the stochastic process whose
unconditional joint probability distribution does not change when shifted in
time. This means that the observations of the time series data have an equal
mean and variance, irrespective of the time period, i.e, they are time
independent.
Since it is an assumption of time series analysis, non-stationary data is often
transformed to become stationary.
The most common cause for violation of the stationarity assumption is trend
in the time series, which can be due to the presence of unit root or a
deterministic trend. In case of the unit root, stochastic shocks have permanent
effects, and the process is not mean-reverting.
In the case of deterministic trend, the process is called a trend stationary
process, and stochastic shocks have only transitory effects after which the
variable tends towards a deterministically evolving(non-constant) mean.
A trend stationary process is not strictly stationary, but can be easily be
transformed into a stationary process by removing the underlying trend,
which is solely a function of time.
One way to make some time series stationary is to compute the differences
between consecutive observations. This is known as differencing.
Transformations such as logarithms can help to stabilize the mean of a time
series by removing the changes in the level of a time series and so
eliminating trend and seasonality.
One of the way for identifying the non-stationary time series is the ACF Plot.
For a stationary process, the ACF will drop to zero relatively quickly, while
the ACF of non-stationary data decreases slowly.

You might also like