Lecture 6 - Time-series Certification Review Exam Questions And Answers (Course) Download To Score A+

time series data - correct answer data collected over time Stationary - correct answer In order to have a linear relationship between Y and X we need all time-series to be stable. We need our null hypothesis to have a unit root. Stationary time-series defined as - correct answer - mean is constant over time - variance is constant over time - the simple correlation between X(t) and X (t-k) depends only on k (the length of the lag) Dicky Fuller Test - correct answer This process examines time-series data on whether it is stationary or non-stationary. If data is non-stationary it converts data to stationary. 1. Tests the null hypothesis if it is stationary and converts non-stationary time-series data into stationary by detrending your variables 2. It tests if it is a unit root Issues with cross-sectional data - correct answer - IID (has independent/individual variables) - heteroscedasticity - too few number of observations Issues with panel data - correct answer - Autocorrelation and have to cluster - Too few numbers of observations Issues with time-series - correct answer - No IID (has no independent/individual variables) - Autocorrelation (even worse) - Stationary Unit Root - correct answer To test: 1. Examine autocorrelations 2. Use Dicky Fuller Test If b1 is close to 1 then it is similar to Random Walk which means it is NON STATIONARY. - If -value > critical values & t-value < critical values -> NO UNIT ROOT = NON-STATIONARY. Non-stationary Process - correct answer the time series possesses a trend - the tendency for the series to either increase or decrease over time. Stationary Process - correct answer the time series varies around some central value and has approximately the same variation over the series Examples of stationary time trends - correct answer - circular time trends - moving average time trends - autoregression time trends Examples of non-stationary time trends - correct answer - linear time trends - exponential time trends - random walk time trends - increasing variance time trends Detrending - correct answer The practice of removing the trend from a time series by fitting a regression line in your data and identify the residuals. If R-squared is high then it is good. What to do when stationary fails - correct answer - transform variables - look at your data, find the cause of the stationary and correct for it - co-integration - just need residuals - dickey fuller test Non-stationary - correct answer When we do not have a linear relationship between Y and X, and all time-series are non-stable. Our null hypothesis is not unit root. Co-integration - correct answer ??? Granger causality - correct answer It is a statistical hypothesis test for determining whether one time-series is useful in forecasting another time-series. Looks if the lags of X is jointly significant to the lags of Y. After running a regression and finding 0 lags --> no change ????