Normally, I used this code below but it is not completely. I have to sum value which start in column 4 to up (some file has value in column 4 to 5 or 4 to 8) but based on column 2 and count in column 2.
awk '{a[$2]+=$4; b[$2]++;} END {for(i in a) print i"\t"a/b;}' inputfile
Thank you very much SriniShoo. It worked :).
However, I have many files which each file contain value column more than 4 to 70 column. I will try to adjust this code for them.