It is perceived that data mining algorithms are used for large sample sets. I am keen to know some references wherein some critical information is given about appropriate sample size for fitting these algorithms not necessarily decision tree algorithms.