The following are the steps to run the program - 

1. Go to "merging_data (folder) -> processing.py"
	Here, processing.py is used to process the data
	i.e. to merge the data given in two files - IF and inputs - according the common name in both.

	Also, processed_data is used to save the processed/merged data from both the files.

2. Then, run "main.py" available in the root folder.
	
	main.py does the following - 
	a) from given 8 inputs, first it creates all the subsets possible and does the following on each subset. 
	b) First it separated the 80% of the processed data to work on.
	c) Then, it creates multi-variate Regression line on the basis of the 80% of the data.
	d) then, it finds the error on the remaining 20% of the data.
	e) It repeats the steps from b) to d) on different 80% of the data and does the same until it found the minimum error possible for the given combination.  
	f) Finally, it prints the minimum error of each subset to the file "output.txt".



OUTPUT - 

The following is the best 10 combinations which can be used to make the regression line with their mean square error and absolute error 

1 = SJR 
2 = H-Index
3 = Total docs (3 yrs)
4 = Total references
5 = Total cites
6 = Citable docs
7 = Cites/Doc
8 = Reference/Doc

combination  ----------  mean absolute error  ------------ mean square error 
error for - 4 5 8  ----  [0.04963302]  ----  [0.01250396]
error for - 4 6 8  ----  [0.0572668]  ----  [0.01621701]
error for - 4 7 8  ----  [0.06233361]  ----  [0.01937769]
error for - 4 5 7  ----  [0.06364095]  ----  [0.02826139]
error for - 4 5 7 8  ----  [0.06442524]  ----  [0.02996358]
error for - 4 7  ----  [0.08782898]  ----  [0.03829524]
error for - 4 8  ----  [0.09597245]  ----  [0.04554588]
error for - 8  ----  [0.10295126]  ----  [0.05241062]
error for - 4 5  ----  [0.15413234]  ----  [0.1177382]
error for - 4 6  ----  [0.18204987]  ----  [0.1639144]