How can we split the huge data into two groups 70% for modelling and 30%for r square, i know a method of choosing randomly 70% from select cases , if that is the way how can we filter and choose the remaining 30% data for ceofficient of determination (use spss program)