Hunt's Algorithm Hunt's algorithm grows a decision tree recursively by partitioning a training data set into smaller, purer subsets. This algorithm contains two steps in order to construct a decision tree.Step 1: which is the terminating step for the recursive algorithm, checks if every record in a node is of the same class. If so, the node is labeled as a leaf node with its classification the class name of all the records within. Step 2: If a node is not pure then selects/creates an attribute test condition to partition the data into two purer data sets. From here a child node is created for each subset.