Is data mining approach a best fit formula for estimation of low-density lipoprotein cholesterol?