Regression typically works best when the ratio of number of data points to number of variables is high. However, in some scenarios, such as clinical trials, the number of data points is limited (given the difficulty in collecting samples from many individuals), and the amount of information collected is high (think of how much information labs give us based on small samples of collected blood).
In these cases, where the ratio of data points to variables is low, one faces difficulty in using the traditional techniques for the following ...
