How does classifiers handle missing data in training?

• Assign most common value of A among other examples sorted to node n
• Assign most common value of A among other examples with same target value
• Assign probability p_i to each possible value v_i of A; Assign fraction p_i of example to each descendant in tree. [1]

Reference:

[1]. http://stackoverflow.com/questions/13425722/how-to-deal-with-missing-attribute-values-in-c4-5-j48-decision-tree