Check out RSS, or use RSS reader to subscribe this item
Confirmation
Authentication email has already been sent, please check your email box: and activate it as soon as possible.
You can login to My Profile and manage your email alerts.
Sponsored by the Center for Science and Technology Development of the Ministry of Education
Supervised by Ministry of Education of the People's Republic of China
By evaluating the information quantity implicated in training samples, the SLRFD algorithm proposed in this paper divides the training set into two parts, confidence domain and non-confidence domain. The confidence domain is used to translate training samples into logical rules and assign class labels to testing samples in the semi-supervised learning framework. Functional dependency rules of probability are deduced based on Armstrong's axioms to serve the purpose of semantics-preserving data dimensionality reduction for classification without sacrificing the ability to discern between samples belonging to different classes. The attributes corresponding to missing values are proved to be redundant, thus the computational complexity while modeling will be reduced. Empirical studies on a set of natural domains show that SLRFD has clear advantages with respect to generalization and probabilistic performance.
Keywords:Computer Science and Technology; Data Mining; Semi-supervised learning; Missing values; Confidence domain; Functional dependency rules of probability