|
|
Sequential Retreating Search Methods in Feature Selection
Somol, Petr ; Pudil, Pavel
Inspired by Floating Search, our new pair of methods, the Sequential Forward Retreating Search (SFRS) and Sequential Backward Retreating Search (SBRS) is exceptionally suitable for Wrapper based feature selection. (Conversely, it cannot be used with monotonic criteria.) Unlike most of other known sub-optimal search methods, both the SFRS and SBRS are parameter-free deterministic sequential procedures that incorporate in the optimization process both the search for the best subset and the determination of the best subset size. The subset yielded by either of the two new methods is to be expected closer to optimum than the best of all subsets yielded in one run of the Floating Search. Retreating Search time complexity is to be expected slightly worse but in the same order of magnitude as that of the Floating Search. In addition to introducing the new methods we provide a testing framework to evaluate them with respect to other existing tools.
|
|
|
Kompozicionální modely domněnkvých funkcí
Jiroušek, Radim ; Vejnarová, Jiřina ; Daniel, Milan
After it has been successfully done in probability and possibility theories, the paper is the first attempt to introduce the operator of composition also for belief functions. We prove that the proposed definition preserves all the necessary properties of the operator enabling us to define compositional models as an efficient tool for multidimensional models representation.
|
|
|
Podmíněná nezávislost ve věrohodnostních funkcích: Příklady
Jiroušek, Radim
The paper presents an additional possibility how to define conditional independence relation for belief functions. The approach is based on the operator of composition originally designed for multidimensional model processing. Not to make confusion with the preceding definitions we call this relation conditional irrelevance. In the paper examples illustrating properties of this relation are presented.
|
|
|
Efektivní algoritmus na hledání redukcí v kompozicionálních modelech
Kratochvíl, Václav
This paper deals with the problem of marginalization of multidimensional probability distributions represented by a compositional model. By the perfect one in this case. From the computational point of view this solution is more efficient than any known marginalization process for Bayesian models. This is because the process mentioned in the paper in a form of an algorithm and takes an advantage of the fact that the perfect sequence models have some information encoded; if can be obtained from the Bayesian networks by an application of rather computationally expensive procedures. One part of that algorithm is marginalization by means of reduction. This paper describe a new faster algorithm to find a reduction in a compositional model.
|
|
|
Má smysl vyvíjet nové metody výběru příznaků?
Somol, Petr ; Novovičová, Jana
One of hot topics discussed recently in relation to pattern recognition techniques is the question of actual performance of modern feature selection methods. Feature selection has been a highly active area of research in recent years due to its potential to improve both the performance and economy of automatic decision systems in various applicational fields, with medical diagnosis being among the most prominent. Feature selection may also improve the performance of classifiers learned from limited data, or contribute to model interpretability. The number of available methods and methodologies has grown rapidly while promising important improvements. Yet recently many authors put this development in question, claiming that simpler older tools show to be actually better than complex modern ones -- which, despite promises, are claimed to actually fail in real-world applications.
|
|
| |
|
|
Rozpoznávání založené na vícerozměrných modelech
Haindl, Michal ; Pudil, Pavel ; Somol, Petr
This chapter explains general model-based approaches to several basic pattern recognition applications followed by a concise description of three fundamental multi-dimensional data model classes. For each model class a solution to parameter estimation and model data synthesis is outlined. Finally an overview of the strengths and weaknesses of studied multi-dimensional data model groups is given.
|
|
|
Texture similarity measure
Vácha, Pavel
This paper surveys the current best texture representations and studies their application for a texture similarity measure development. A simple experiment that evaluates texture similarity is proposed and the performance of several most advanced texture features is verified on it. In order to eliminate the influence of spectral information monospectral textures are considered in this study only. The paper suggests the L1 norm with either Markovian or Gabor features as the best texture similarity measure.
|
|
| |
|
| |
guest ::
