Data section for draft.txt

-The raw data used in this analysis are daily closing prices of the *equally weighted{market cap?}* FTSE/JSE Top40 index for the period dd/mm/yyyy to dd/mm/yyyy.
-The choice to use a post-crisis sampling period is consistent with the methodology employed by *CITE: Jiang et el(2016)* in analysing analysing
stock markets after the financial crisis . Furthermore, the choice of the starting date is consistent with what the International Monetary Fund(IMF) considers to be the post-crisis period (Strauss- see proposal). 

-The data was transformed into returns and log returns for forecasting purposes since
		- stock price returns are generally considered independent
		- a means of preventing issues with spurious regression *cite*
		- noise
		- standard practice
-We consider both returns and log returns as a means of comparing if the nature of the transformation can influence the comparative performance of the forecasting mmodels
The returns data will be denoted by __ and were evaluated using the formula ____ whilst log returns will be denoted by ____ and were evaluated using the formula ____ .

- The data, which was sourced from Datastream, has no missing values but excludes non-trading days such as weekends and holidays. However, it was cleaned to remove duplicates and identify outliers. 
it was necessary that the data be cleaned to identify the presence of outliers and