How to impute missing value using mean , median and mode for labelled dataset ?

I am doing missing value imputation using mean , median and mode . I have 150*10 dimensioned dataset and the last attribute has labels so the actual data is 150*9 but i don't know how to get mean values for each different labels. my dataset has three labels first 50 has 1 label the second 50 has 2 label and the third 50 has 3 label . and How to impute this values in the place of missing value?