CREATING METHOD FOR DECISION TREE USING TIME-WEIGHTED ENTROPY AND RECORDING MEDIUM THEREOF

摘要

A method for generating a decision tree using time-weighted entropy and a recording medium thereof are provided to predict activity of a user or recognize a tendency of the user by analyzing preference data of the user in a ubiquitous environment. A test for selecting an attribute to be a node of a current level among the attributes of a plurality of instance data by using the time-weighted entropy is performed(ST2010). The attribute having the highest information quantity obtained from the test is selected as the node(ST2033). Previous two steps are repeatedly performed until all nodes of the current level are selected(ST2034). If all nodes of the current level are selected, all previous steps are repeatedly performed for the level just below the current level(ST2034). The decision tree generated by the previous two steps is generated until all low instance data groups are unified into one class and system entropy gets to be zero(ST2050). The time-weighted entropy of the recent tendency has bigger weight.