You are given time series of rainfall data for different meteorological
divisions of India. Data are from monsoon month and monsoon season and annual
dataset. Each student needs to use only one dataset. For example, Orissa for July
month or Annual rainfall for Orissa. Each student name is given in the excel sheet
and a column of data is assigned. You need to perform linear and Mann Kendall
trend analysis. Decompose the time series into different components. Final step is
to do forecasting using ARIMA. You need to submit all the scripts for your analysis.
1. How different are linear and monotonic trend for the time series you are
working on? Provide interpretation of the result. (5)
2. Provide the steps of decomposition. Which model did you use and why?
Provide the graph(s) in support of your answer. Show all the components of
the decomposition and describe them. (5)
3. Comment on stationarity of time series. Provide any plot required to
illustrate the stationarity/non-stationarity of time series. (5)
4. Comment on the ARIMA model you used for the time series forecasting. How
did you find out best model? Describe it with ACF and PACF graph. (10 pts).
5. Provide the graph of training and test data along with predicted values (1 pt).
0 comments:
Post a Comment