Digital printing has begun to play a more significant role in the current commercial printing market, and the demand for accurate color reproduction has also increased. Nonetheless, achieving wide selection of printable substrates imposes a limiting factor on color accuracy for a printing press manufacturer. The compromised solution is to only provide and maintain certain generic substrate ICC profiles based on the physical features of substrates. The main concern for this approach is that it is very difficult to establish a correlation between each substrate characteristics and the chosen imaging process. We propose a nonparametric approach by first sifting through a representative set of substrate-specific ICC profiles. Design a feature dimension reduction algorithm based on the independent component analysis, and we will be able to identify a small set of substrates achieving the optimal generic ICC profile performance against investigated substrates. At last, we will identify the important physical characteristics of a substrate affecting the performance of the ICC profile with respect to the selected imaging process.