Relational interpretations of neighborhood operators and rough set approximation operators
Information Sciences—Informatics and Computer Science: An International Journal
Rough Sets: Theoretical Aspects of Reasoning about Data
Rough Sets: Theoretical Aspects of Reasoning about Data
A Generalized Definition of Rough Approximations Based on Similarity
IEEE Transactions on Knowledge and Data Engineering
On the Extension of Rough Sets under Incomplete Information
RSFDGrC '99 Proceedings of the 7th International Workshop on New Directions in Rough Sets, Data Mining, and Granular-Soft Computing
On the Unknown Attribute Values in Learning from Examples
ISMIS '91 Proceedings of the 6th International Symposium on Methodologies for Intelligent Systems
A Two-Phase Model for Learning Rules from Incomplete Data
Fundamenta Informaticae - Fundamentals of Knowledge Technology
Two-phase rule induction from incomplete data
RSKT'08 Proceedings of the 3rd international conference on Rough sets and knowledge technology
Positive approximation and converse approximation in interval-valued fuzzy rough sets
Information Sciences: an International Journal
Hybridization of rough sets and statistical learning theory
Transactions on rough sets XIII
Rule extraction based on granulation order in interval-valued fuzzy information system
Expert Systems with Applications: An International Journal
Generalized rough sets and implication lattices
Transactions on rough sets XIV
An interval set model for learning rules from incomplete information table
International Journal of Approximate Reasoning
A Two-Phase Model for Learning Rules from Incomplete Data
Fundamenta Informaticae - Fundamentals of Knowledge Technology
Rough set based pose invariant face recognition with mug shot images
Journal of Intelligent & Fuzzy Systems: Applications in Engineering and Technology
| Hi-index | 0.00 |
In this paper we present results of experiments conducted to compare three types of missing attribute values: lost values, "do not care" conditions and attribute-concept values. For our experiments we selected six well known data sets. For every data set we created 30 new data sets replacing specified values by three different types of missing attribute values, starting from 10%, ending with 100%, with increment of 10%. For all concepts of every data set concept lower and upper approximations were computed. Error rates were evaluated using ten-fold cross validation. Overall, interpreting missing attribute values as lost provides the best result for most incomplete data sets.