Feature Selection Using Typical ε: Testors, Working on Dynamical Data
Typical e:testors are useful to do feature selection in supervised classification problems with mixed incomplete data, where similarity function is not the total coincidence, but it is a one threshold function. In this kind of problems, modifications on the training matrix can appear very frequently. Any modification of the training matrix can change the set of all typical ε:testors, so this set must be recomputed after each modification. But, complexity of algorithms for calculating all typical ε:testors of a training matrix is too high. In this paper we analyze how the set of all typical ε:testors changes after modifications. An alternative method to calculate all typical ε:testors of the modified training matrix is exposed. The new method’s complexity is analyzed and some experimental results are shown.
- 4.Dmitriev, A.N., Zhuravlev, Y.I., Krendeliev, F.P.: About mathematical principles of objects and phenomena classification. Diskretni Analiz 7, 3–15 (1966) (in Russian)Google Scholar
- 5.Carrasco-Ochoa, J.A., Ruiz-Shulcloper, J.: Sensitivity Problems of the Set of Typical Testors of a Boolean Matrix”. In: Proceedings of the III Iberoamerican Symposium on Pattern Recognition. SIARP 1998, Mexico, pp. 257–266 (1998) (in Spanish)Google Scholar
- 6.Carrasco-Ochoa, J.A.: Sensitivity in Logical Combinatorial Pattern Recognition. PhD Thesis, CIC-IPN, Mexico (2001)Google Scholar