Approximate Algorithm for Minimization of Decision Tree Depth

In the paper a greedy algorithm for minimization of decision tree depth is described and bounds on the algorithm precision are considered. This algorithm is adapted for application to data tables with both discrete and continuous variables, which can have missing values. To this end we transform given data table into a decision table. Under some natural assumption on the class N P the considered algorithm is close to unimprovable approximate polynomial algorithms for minimization of decision tree depth.