The concept of `diversity' has been one of the main open issues in the field of multiple classifier systems. In this paper we address a facet of diversity related to its effectiveness for ensemble construction, namely, explicitly using diversity measures for ensemble construction techniques based on the kind of overproduce and choose strategy known as ensemble pruning. Such a strategy consists of selecting the (hopefully) more accurate subset of classifiers out of an original, larger ensemble. Whereas several existing pruning methods use some combination of individual classifiers' accuracy and diversity, it is still unclear whether such an evaluation function is better than the bare estimate of ensemble accuracy. We empirically investigate this issue by comparing two evaluation functions in the context of ensemble pruning: the estimate of ensemble accuracy, and its linear combination with several well-known diversity measures. This can also be viewed as using diversity as a regularizer, as suggested by some authors. To this aim we use a pruning method based on forward selection, since it allows a direct comparison between different evaluation functions. Experiments on thirty-seven benchmark data sets, four diversity measures and three base classifiers provide evidence that using diversity measures for ensemble pruning can be advantageous over using only ensemble accuracy, and that diversity measures can act as regularizers in this context.
In this paper we propose cryptosystems based on subgroup distortion in hyperbolic groups. We also include concrete examples of hyperbolic groups as possible platforms.
A class of Xorshift Random Number Generators (RNGs) are introduced by Marsaglia. We have proposed an algorithm which constructs a primitive Xorshift RNG from a given prim- itive polynomial. We also have shown a weakness present in those RNGs and suggested its solution. A separate algorithm also proposed which returns a full periodic Xorshift generator with desired number of Xorshift operations.
The paper presents a Car Sequencing Problem, widely considered in the literature. The issue considered by the researchers is only a reduced problem in comparison with the problem in real automotive production. Consequently, a newconcept, called Paint Shop 4.0., is considered from the viewpoint of a sequencing problem. The paper is a part of the previously conducted research, identified as Car Sequencing Problem with Buffers (CSPwB), which extended the original problem to a problem in a production line equipped with buffers. The new innovative approach is based on the ideas of Industry 4.0 and the buffer management system. In the paper, sequencing algorithms that have been developed so far are discussed. The original Follow-up Sequencing Algorithm is presented, which is still developed by the authors. The main goal of the research is to find the most effective algorithm in terms of minimization of painting gun changeovers and synchronization necessary color changes with periodic gun cleanings. Carried out research shows that the most advanced algorithm proposed by the authors outperforms other tested methods, so it is promising to be used in the automotive industry.
Although the explicit commutativitiy conditions for second-order linear time-varying systems have been appeared in some literature, these are all for initially relaxed systems. This paper presents explicit necessary and sufficient commutativity conditions for commutativity of second-order linear time-varying systems with non-zero initial conditions. It has appeared interesting that the second requirement for the commutativity of non-relaxed systems plays an important role on the commutativity conditions when non-zero initial conditions exist. Another highlight is that the commutativity of switched systems is considered and spoiling of commutativity at the switching instants is illustrated for the first time. The simulation results support the theory developed in the paper.
The problem of optimally controlling a Wiener process until it leaves an interval (a; b) for the first time is considered in the case when the infinitesimal parameters of the process are random. When a = ��1, the exact optimal control is derived by solving the appropriate system of differential equations, whereas a very precise approximate solution in the form of a polynomial is obtained in the two-barrier case.