Explaining prediction models and individual predictions with feature contributions

We present a sensitivity analysis-based method for explaining prediction models that can be applied to any type of classification or regression model. Its advantage over existing general methods is that all subsets of input features are perturbed, so interactions and redundancies between features are taken into account. Furthermore, when explaining an additive model, the method is equivalent to commonly used additive model-specific methods. We illustrate the method’s usefulness with examples from artificial and real-world data sets and an empirical analysis of running times. Results from a controlled experiment with 122 participants suggest that the method’s explanations improved the participants’ understanding of the model.     

Porous Perovskite Fibers – Preparation by Wet Phase Inversion Spinning and Catalytic Activity

Porous perovskite (LaMnO₃) fibers were prepared by means of wet phase inversion spinning. The influence of different spinning procedures, slurry and coagulation bath composition on fiber shape and pore morphology was studied. The catalytic activity of the prepared fibers was tested for carbon monoxide oxidation as a model reaction in a differential recycle reactor. The results revealed that by suitable choice of process conditions porous catalytically active fibers can be prepared. Catalytic measurements confirmed that the catalytic fibers exhibit an open structure that allows full utilization of the catalytically active surface without intraparticle diffusional limitations.     

Thermal Expansion,Elastic and Magnetic Properties of FeCoNiCu-Based High-Entropy Alloys Using First-Principle Theory

The effects of V, Cr, and Mn on the magnetic, elastic, and thermal properties of FeCoNiCu high-entropy alloy are studied by using the exact muffin-tin orbitals method in combination with the coherent potential approximation. The calculated lattice parameters and Curie temperatures in the face-centered-cubic structure are in line with the available experimental and theoretical data. A significant change in the magnetic behavior is revealed when adding equimolar V, Cr, and Mn to the host composition. The three independent single-crystal elastic constants are computed using a finite strain technique, and the polycrystalline elasticity parameters including shear modulus, Young’s modulus, Pugh ratio, Poisson’s ratio, and elastic anisotropy are derived and discussed. The effects of temperature on the structural parameters are determined by making use of the Debye–Grüneisen model. It is found that FeCoNiCuCr possesses a slightly larger thermal expansion coefficient than do the other alloys considered here.     

Crosslinking of polybenzimidazole membranes by divinylsulfone post‐treatment for high‐temperature proton exchange membrane fuel cell applications

Phosphoric acid‐doped polybenzimidazole (PBI) has been suggested as a promising electrolyte for proton exchange membrane fuel cells operating at temperatures up to 200 °C. This paper describes the development of a crosslinking procedure for PBI membranes by post‐treatment with divinylsulfone. The crosslinking chemistry was studied and optimized on a low‐molecular‐weight model system and the results were used to optimize the crosslinking conditions of PBI membranes. The crosslinked membranes were characterized with respect to chemical and physiochemical properties, showing improved mechanical strength and oxidative stability compared with their linear analogues. Fuel cell tests were further conducted in order to demonstrate the feasibility of the crosslinked membranes. Copyright © 2011 Society of Chemical Industry     

From Side Chains Rattling on Picoseconds to Ensemble Simulations of Protein Folding

Simulations of biological macromolecules have evolved tremendously since the discoveries of the 1970s. The field has moved from simple simulations in vacuo on picosecond scales to milliseconds of accurate sampling of large proteins, and it has become a standard tool in biochemistry and biophysics, rather than a dedicated theoretical one. This is partly due to increasing computational power, but it would not have been possible without huge research efforts invested in new algorithms and software. Here, we illustrate some of this development, both past and future challenges, and in particular, discuss how the recent introduction of modern ensemble methods is breaking the trend of ever-longer simulations to instead focus on throughput and sampling. This has not only helped simulations become much more accurate, but it provides statistical error estimates, which are critical, as simulations are increasingly used to predict properties that have not yet been measured experimentally.     

Two-way Bayesian hierarchical phylogenetic models: An application to the co-evolution of gp120 and gp41 during and after enfuvirtide treatment

Enfuvirtide (ENF) is a fusion inhibitor that prevents the entry of HIV virions into target cells. Studying the characteristics of viral evolution during treatment and after a treatment interruption can lend insight into the mechanisms of viral evolution and fitness. Although interruption of anti-HIV therapy often results in rapid emergence of an archived “wild-type” virus population, previous work from our group indicates that when only ENF is interrupted, viral gp41 continues to evolve forward and resistance mutations are lost due to back-mutation and remodeling of the envelope protein. To examine the co-evolution of gp120 and gp41 during ENF interruption we extend the Bayesian Hierarchical Phylogenetic model (HPM). Current HPMs enforce conditional independence across all outcomes while biologically all gene regions within a patient should return the same tree unless recombination confers an evolutionary selective advantage. A two-way-interaction HPM is proposed that provides middle ground between these two extremes and allows us to test for differences in evolutionary pressures across gene regions in multiple patients simultaneously. When the model is applied to a well-characterized cohort of HIV-infected patients interrupting ENF we find that across patients, the virus continued to evolve forward in both gene regions. Overall, the hypothesis of independence over dependence between the gene regions is supported. Models that allow for the examination of co-evolution over time will be increasingly important as more therapeutic classes are developed, each of which may impact other through novel and complex mechanisms.     

n−3 fatty acids,heart rate variability,and sudden cardiac death

Sudden cardiac death (SCD) is a major cause of mortality in Western countries. Furthermore, SCD is often the first manifestation of coronary artery disease, making it difficult to prevent. Heart rate variability (HRV), which can be determined by extended recording of the heart rate by 24-h Holter monitoring, has been shown to be one of the best predictors of the risk of SCD. There is increasing evidence from animal experiments and clinical trials in humans that n?3 fatty acids reduce the risk of SCD. We have studied the effect of n?3 fatty acids on HRV and present data clearly showing that n?3 fatty acids increase HRV. This adds further to the hypothesis that an increased intake of n?3 fatty acids may reduce the risk of SCD.     

Marine n-3 fatty acids: Basic features and background

There is some evidence from epidemiology that intake of n-3 polyunsaturated fatty acids (PUFA) from seafood may protect against coronary artery disease (CAD). This hypothesis is further supported from animal data showing a beneficial effect of n-3 PUFA on thrombosis and atherosclerosis in animals fed fish oils in most, but not all, studies. There are several mechanisms by which an increased intake of marine n-3 PUFA may protect against CAD; the most universal finding is a reduction of plasma triglycerides. It is puzzling, however, that a very low amount of n-3 PUFA, with no known beneficial biochemical effects, seems to be cardioprotective. It has therefore been of paramount interest to perform clinical trials. Such evidence and trials are discussed in later chapters, and the results have been very encouraging.     

Testing probability matching and episodic retrieval accounts of response repetition effects in task switching.

This study takes inventory of available evidence on response repetition (RR) effects in task switching, in particular the evidence for RR cost when the task switches. The review reveals that relatively few task-switching studies in which RR effects were addressed have shown statistical support for RR cost, and that almost all are affected by 1 of 2 potential artifacts, either a response bias caused by disallowing stimulus repetitions or the effect of including stimulus repetitions in data analysis. New results with these factors controlled support an episodic retrieval model in which features of the retrieved trace, including the stimulus but also the task, task cue, and response, facilitate or interfere with performance depending on whether they match or mismatch the current processing context. (PsycINFO Database Record (c) 2011 APA, all rights reserved)