Thesis supervisor: József Dombi
Location of studies (in Hungarian): SZTE Abbreviation of location of studies: SZTE
Description of the research topic:
Decision trees are of great importance among learning algorithms, as the black box character does not exist in the this area, i.e. the result can be interpreted after learning. These types of learning algorithms are of great importance in the field of deep neural networks. Classical decision trees are interpreted for discrete variables i.e. categories. The subject of the present research is the establishing and application of continuous variables in decision trees. These structures can be defined with the so-called soft inequality (sigmoid function). Using the sigmoid function the decision tree can be constructed. The resulting structures can be considered as special neural networks and their interpretation can be given. The task is to create and study these structures.
Required language skills: English Number of students who can be accepted: 2