Evolutionary model trees for handling continuous classes in machine learning

Model trees are a particular case of decision trees employed to solve regression problems. They have the advantage of presenting an interpretable output, helping the end-user to get more confidence in the prediction and providing the basis for the end-user to have new insight about the data, confirming or rejecting hypotheses previously formed. Moreover, model trees present an acceptable level of predictive performance in comparison to most techniques used for solving regression problems. Since generating the optimal model tree is an NP-Complete problem, traditional model tree induction algorithms make use of a greedy top-down divide-and-conquer strategy, which may not converge to the global optimal solution. In this paper, we propose a novel algorithm based on the use of the evolutionary algorithms paradigm as an alternate heuristic to generate model trees in order to improve the convergence to globally near-optimal solutions. We call our new approach evolutionary model tree induction (E-Motion). We test its predictive performance using public UCI data sets, and we compare the results to traditional greedy regression/model trees induction algorithms, as well as to other evolutionary approaches. Results show that our method presents a good trade-off between predictive performance and model comprehensibility, which may be crucial in many machine learning applications. (C) 2010 Elsevier Inc. All rights reserved.

Selective Wrapping and Supramolecular Structures of Polyfluorene-Carbon Nanotube Hybrids

We report on the photophysical properties of single-walled carbon nanotube (SWNT) suspensions In toluene solutions of poly[9,9-dioctylfluorenyl-2,7-diyl](PFO). Steady-state and time-resolved photoluminescence spectroscopy in the near-infrared and visible spectral regions are used to study the interaction of the dispersed SWNTs with the wrapped polymer. Molecular dynamics simulations of the PFO-SWNT hybrids in toluene were carried out to evaluate the energetics of different wrapping geometries. The simulated fluorescence spectra in the visible region were obtained by the quantum chemical ZINDO-CI method, using a sampling of structures obtained from the dynamics trajectories. The tested schemes consider polymer chains aligned along the nanotube axis, where chirality has a minimal effect, or forming helical structures, where a preference for high chiral angles is evidenced. Moreover, toluene affects the polymer structure favoring the helical conformation. Simulations show that the most stable hybrid system is the PFO-wrapped (8,6) nanotube, in agreement with the experimentally observed selectivity.

Prigogine-Defay Ratio for an Ionic Glass-Former: Molecular Dynamics Simulations

The pressure dependence of the glass-transition temperature, T(g)(P), of the ionic glass-former 2Ca(NO(3))(2) center dot 3KNO(3), CKN, has been obtained by molecular dynamics (MD) simulations The liquid-glass difference of thermal expansivity, Delta alpha, heat capacity, Delta C(p), and isothermal compressibility, Delta kappa, have been calculated as a function of pressure. It has been found that the Ehrenfest relation dT(g)/dP = TV Delta alpha/Delta C(p) predicts the pressure dependence of T, but the other Ehrenfest relation, dT(g)/dP = Delta kappa/Delta alpha, does not. Consequently, the Prigogine-Defay ratio, Pi = Delta C(p)Delta kappa/TV Delta alpha(2), is Pi similar to 1.2 at low pressures, but increases 1 order of magnitude at high pressures. The pressure dependence of the Prigogine-Defay ratio is interpreted in light of recent explanations for the finding Pi > 1.