Separable Data Aggregation by Layers of Binary Classifiers
Aggregating layers can be designed from binary classifiers on the principle of preserving data sets separability. Formal neurons or logical elements are treated here as basic examples of binary classifiers. Learning data sets are composed of such feature vectors which are linked to particular categories (classes). Separability of the learning sets is preserved during transformation of feature vectors from these sets by a dipolar layer of binary classifiers. The dipolar layer separates all such pairs of feature vectors that have been linked to different classes and belong to different learning sets.
KeywordsFeature vectors Binary classifiers Separable data aggregation Dipolar aggregation Hierarchical networks Deep learning
The presented study was supported by the grant S/WI/2/2013 from Bialystok University of Technology and funded from the resources for research by Polish Ministry of Science and Higher Education.
- 1.Hand, D., Smyth, P., Mannila, H.: Principles of Data Mining. MIT Press, Cambridge (2001)Google Scholar
- 3.Bobrowski, L.: Data mining based on convex and piecewise linear criterion functions. Technical University Białystok (2005). (in Polish)Google Scholar
- 4.Bobrowski, L.: Piecewise-linear classifiers, formal neurons and separability of the learning sets. In: Proceedings of ICPR 1996, pp. 224–228, 13th International Conference on Pattern Recognition, Vienna, Austria, 25–29 August 1996 (1996)Google Scholar
- 5.Bobrowski, L., Topczewska, M.: Dipolar data aggregation in the context of deep learning. In: Kůrková, V., Manolopoulos, Y., Hammer, B., Iliadis, L., Maglogiannis, I. (eds.) ICANN 2018. LNCS, vol. 11141, pp. 574–583. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-01424-7_56CrossRefGoogle Scholar
- 7.Arel, I., Rose, D.C., Karnowski, T.P.: Deep machine learning – a new frontier in artificial intelligence – a survey paper. IEEE Comput. Intell. Mag. (2013)Google Scholar
- 10.Dua, D., Karra Taniskidou, E.: UCI Machine Learning Repository. University of California, School of Information and Computer Science, Irvine, CA (2017). http://archive.ics.uci.edu/ml