# TIME SERIES ANALYSIS - REQUEST

Cancelled

## Job Description

My Greetings ...

Kindly, I have 4 folders of the name (S & N),(AT & N),(P &N),(R &N)
These folders contain Excel files of kind A,D,E.
The number of files of kind A is 74 files.
The number of files of kind D is 74 files.
The number of files of kind E is 74 files.
The files of the kind A & E : each file consists of 12 tables with span start from 1950-2011
The files of kind D : each file consists of 4 tables with span start from 1950-2011
The variables of symbols (S,AT,P) represent the Dependent
The variables of symbols (N) ,represent the Independent.
My request is to calculate the
Trend Equation
Correlation Coefficient
between the dependent variables (S,AT,P,R) and (N).
Note:
-The first folder (S & N) -It must be linearly De Trend -Because the Trend is Combined=Deterministic+Stochastic
-The folders (AT&N),(P&N),(R&N)- the Trend is Deterministic .You are free.

The data is TIME SERIES and each time series consists of 4 components as you know

Time Series= Trend*Seasonality*Periodicity*Irregular
When you plot each table you will see these components.The good job is how to capture the TREND line.

The correct Trend and correct Correlation Coefficient satisfied if you proceed
the following tests;
1-Autocorrelation
2-Hetereoscedasticity
3-T-test
4-F-tests
5-Sig less 0.05 for Correlation coefficients
6-Sig less 0.05 for Trend constants

All the time series have the autocorrelation problem and Hetereoscedasiticity problem so the good method is generalized Least Squares (GLS) -STATA Package-to capture the linear TREND.
Or another method as ARIMA ,You are free.
So,you have to remove the autocorrelation from the Dependent time series then you have to correlate with independent time series to calculate the correlation coefficients.

Please, If you have time and ability please inform me to send you the data after the convention between us for the cost and time

Regards
ALI
Ph.D. Student
University of Szczecin
Poland
Mob: 0048 535 240 762762

Open Attachment