is a predicted value of y, based on the least square model fit, and adjusted R2=1???(1???R2)*(n???1)/(n???1???c), where n is a number of observations, and c is a number of coefficients estimated; and using the value 100* R2 as the similarity measure between the two data sets. 28.