Training set selection for SVMs
Evolutionary methods for support vector machines training set optimization project is supported by the Polish Ministry of Science and Higher Education
under research grant no. IP2012 026372 from the Science Budget 2013-2015. The project involves the following tasks:
- create the validation procedure based on benchmark data sets,
- implement and validate the state-of-the-art methods for training set selection,
- design memetic methods which exploit the knowledge attained during the evolutionary optimization to guide the search process,
- design methods for local optimization,
- analyze the possibilities of applying the proposed techniques to select an optimal training set from unlabeled data.
Related publications:
-
Nalepa, J., Kawulok, M.:
Adaptive memetic algorithm enhanced with data geometry analysis to select training data for SVMs,
Neurocomputing, vol. 185, pp. 113--132, 2016.
-
Kawulok, M., Nalepa, J.:
Towards robust SVM training from weakly labeled large data
sets, in Proc. Asian Conference on Pattern Recognition, ACPR 2015 /in print/.
-
Nalepa, J., Siminski, K., Kawulok, M.:
Towards parameter-less support vector machines, in Proc. Asian Conference on Pattern Recognition, ACPR 2015 /in print/.
-
Nalepa, J., Szymanek, J., Kawulok, M.:
Real-time people counting from depth images.
In Kozielski, S., Mrozek, D., Kasprowski, P., Malysiak-Mrozek, B.,
Kostrzewa, D., eds.: Beyond Databases, Architectures and Structures. Volume
521 of Communications in Computer and Information Science.
Springer International Publishing (2015) 387-397
-
Kawulok, M., Nalepa, J.:
Dynamically adaptive genetic algorithm to select training data for
{SVMs}.
In Bazzan, A.L., Pichara, K., eds.: Advances in Artificial
Intelligence -- IBERAMIA 2014. Volume 8864 of Lecture Notes in Computer
Science.
Springer International Publishing (2014) 242-254
-
Nalepa, J., Kawulok, M.:
A memetic algorithm to select training data for support vector
machines.
In: Proc. Conference on Genetic and Evolutionary Computation. GECCO
'14, New York, NY, USA, ACM (2014) 573-580
-
Kawulok, M., Nalepa, J.:
Hand pose estimation using support vector machines with evolutionary
training.
In: Proc. International Conference on Systems, Signals and Image
Processing (IWSSIP 2014). (May 2014) 87-90
-
Nalepa, J., Kawulok, M.:
Adaptive genetic algorithm to select training data for support vector
machines.
In Esparcia-Alcazar, A.I., Mora, A.M., eds.: Applications of
Evolutionary Computation. Lecture Notes in Computer Science.
Springer Berlin Heidelberg (2014) 514-525
-
Kawulok, M., Nalepa, J.: Support vector machines training data selection using
a genetic algorithm, in Statistical Techniques in Pattern Recognition, S+SSPR
2012, vol. 7626 of LNCS, pp. 557-565, Springer-Verlag Berlin Heidelberg, 2012.
Our data sets
The project involved creation of several data sets (real-life and artificial 2D points) which are available here.