No. 1/2015

Online archive of Computer Science and Mathematical Modelling

No. 1/2015

  1. A. Ameljańczyk - Pareto filter in the process of multi-label classifier synthesis in medical diagnostics support algorithms
    Pages: 5 - 10
    Abstract: The paper presents the possibility of using multi-criteria optimization methods for simple classifiers fusion in a more precise and reliable classifiers complex. There are defined simple classifiers (one label) in the form of classifier committees and sample the synthesis relations of classifiers allow to obtain classifiers filed with improved properties.
    Keywords Pareto filter, medical diagnostics, the committee classifiers, classifiers synthesis, not dominated diagnosis
    Full article: PDF icon 5_10_aameljanczyk_csmm_1_2015.pdf
  2. A. Ameljańczyk - Property analysis of multilabel classifiers in the example of determining the initial medical diagnosis
    Pages: 11 - 16
    Abstract: In the paper analysis of properties of medical diagnoses acquired using simple and complex classifiers was performed. The introduced terms are illustrated with a comprehensive example in the field of medical diagnostics.
    Keywords medical diagnostics, committee classifiers, synthesis of classifiers, not dominated diagnosis classifier reliability
    Full article: PDF icon 11_16_aameljanczyk_csmm_1_2015.pdf
  3. A. Chojnacki, G. Pieniążek - Time based deterministic model of information security of a business organization
    Pages: 17 - 24
    Abstract: This paper discusses information security of a business organization using a time-based mathematical deterministic model. The model addresses key features of a business organization from the point of view of information security and calculates the level of information security based on quantitative values. Next, the introduced model is used to evaluate the level of information security that could be achieved for known threats within a given budget. For this reason, an optimization problem of safeguard implementation is formulated and an optimization method based on dynamic programming is used to address the issue. Two samples, local and global security metrics, defined in the model are described and one of them is used in optimizing safeguard implementation.
    Keywords information security, deterministic model, dynamic programming
    Full article: PDF icon 17_24_achojnacki_gpieniazek_csmm_1_2015.pdf
  4. M. Chudy - Model of attrition process control
    Pages: 25 - 30
    Abstract: The essence of the presented attrition process control relies on the solving defined sequence of target assignment problems at specific moments. The sequence of these moments is obtained for both sides of the battle. The model takes into account the changes of the number of means and targets as well as changes in environmental conditions. It is reflected in the parameters of problems. Each of the considered assignment problems belongs to the class of general assignment problems which does not contain totally unimodular matrix factors.
    Keywords mathematical modeling, attrition process, assignment problem
    Full article: PDF icon 25_30_mchudy_csmm_1_2015.pdf
  5. T. Rzeźniczak - Evolutionary algorithms for Map of Attributes optimization
    Pages: 31 - 43
    Abstract: Map of Attributes (MoA) is a visualization technique that allows to construct graphical representation of abstract entities. The technique is intended to aid recognition of the entities’ representations through the effective use of human perception abilities. A certain difficulty in the application of MoA is the computational complexity of finding an optimal map. The study presents a heuristic approach, based on evolutionary algorithms (EA), to constructing MoA visualization. The method was evaluated using the repository of disease entities as an input dataset. Several different setups of EA were tested; these were configurations with well-known evolution operators, as well as setups with newly proposed operators for the matrix representation of chromosome. Detailed results and analysis of conducted experiments are presented.
    Keywords evolutionary algorithms, data visualization, evolution operators
    Full article: PDF icon 31_43_trzezniczak_csmm_1_2015.pdf