Research Article

Distance Variance Score: An Efficient Feature Selection Method in Text Classification

Algorithm 1

Algorithm of feature selection using DVS.
Procedure  
begin
let   denotes the map storing and () and
 # is the matrix to store value of city block distance among documents and
 #
 #when ,
for each in   do
  for each in   do
   ;
  end
end
 #calculate for each
for   in   do
  ;
  ;
end
 #sort in decreasing order according to value
;
 #select and return the first features
return  ;
end
Procedure  
# is the matrix to store value of city block distance among documents and
#
#when ,
begin
let   denotes the standard variance of
let   denotes feature distance contributions of
let   denotes the index vector of samples in and
 #calculate .  See Section 3.1
for each in   do
  for each in and   do
   #calculate city block distance between and . See Section 3.1
   ;
   #calculate city block distance between and . See Section 3.1
   ;
    += ;
  end
end
return  ;
end