Applied Time Series Analysis Programming assignment No. 1
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
Applied Time Series Analysis
Programming assignment No. 1
Tasks:
1) Download the time series for quarterly seasonally adjusted German real GDP from the FRED database of the St. Louis Fed and save it as a *.csv file (https://fred.stlouisfed. org/series/CLVMNACSCAB1GQDE). Load the data into Python and provide a time series plot of the raw series.
2) Write a function that performs some standard transformations (i.e. log, first differences of logs, first differences, seasonal differences, seasonal differences of logs) on a user-specified time series. Using the GDP data from 1), create plots for the quarterly and yearly growth rates of real GDP.
3) Write a function that computes for any time series input the sample autocorrelations and sample partial autocorrelation up to some user-specified lag h. Program a plot of the sample autocorrelations and sample partial autocorrelation for the quarterly and the yearly growth rate of German real GDP. Compare the four plots and explain.
4) Create a function to implement the Hodrick-Prescott(HP)-filter for the log of real German GDP (Hint: Results from problem set 1 may be useful). Provide a plot of the cyclical com- ponent from the HP-filter.
5) Create a function that generates artificial data of an MA(1) process using the data generating
process
for t = 2,...,T, where εt is a normally distributed white noise process with mean 0 and
variance σ2 = 1. The function should take as an input the parameter θ, the time series length T and should return a vector of observations on yt .
Generate two time series for θ = 0.3 and θ = −0.5 with T = 500 observations according to equation (1) and compute the sample autocorrelation and sample partial autocorrelation functions for both time series. Compare your results with the expected patterns of the theo- retical ACFs and PACFs of the processes.
2022-05-11