The construction of efficient and effective decision trees remains a key topic in machine learning because of their simplicity and flexibility. A lot of heuristic algorithms have been proposed to construct near-optimal decision trees. Most of them, however, are greedy algorithms which have the drawback of obtaining only local optimums. Besides, common split criteria, e.g. Shannon entropy, Gain ...

We revisit the issues on the thermodynamic property of stellar self-gravitating system arising from Tsallis’ non-extensive entropy. Previous papers (Taruya & Sakagami, Physica A 307 (2002) 185 (cond-mat/0107494); ibid. (2002) in press (condmat/0204315)) have revealed that the extremum-state of Tsallis entropy characterized by the so-called stellar polytrope has consistent thermodynamic structur...

We present a first physical application of Tsallis’ generalized entropy to the thermodynamics of self-gravitating systems. The stellar system confined in a spherical cavity of radius re exhibits an instability, so-called gravothermal catastrophe, which has been originally investigated by Antonov (Vest.Leningrad Gros.Univ. 7 (1962) 135) and Lynden-Bell & Wood (Mon.Not.R.Astron.Soc. 138 (1968) 49...

1. Introduction This paper reviews recent work on the properties of Tsallis distributions, Tsallis statistics, anomalous diffusion and multi-fractal processes. Processes of this kind have the capacity to represent natural phenomena that are associated with turbulence and emerging complexity. However, in addition they may also capture crucial aspects of human decision-making in response to condi...

We address this work to investigate symbolic sequences with long-range correlations by using computational simulation. We analyze sequences with two, three and four symbols that could be repeated l times, with the probability distribution p(l) ∝ 1/lμ. For these sequences, we verified that the usual entropy increases more slowly when the symbols are correlated and the Tsallis entropy exhibits, f...

We show that there is an effect of nonextensivity acting upon the BCS model for superconductors in the ground state that motivates its study in the Tsallis’ statistical framework. We show that the weak-coupling limit superconductors are well described by q ∼ 1, where q is a real parameter which characterizes the degree of nonextensivity of the Tsallis’ entropy. Nevertheless, small deviations wi...