A Meta-analysis on Classification Model Performance in Real-World Datasets: An Exploratory View

The No Free Lunch (NFL) Theorem imposes a theoretical restriction on optimization algorithms and their equal average performance on different problems, under some particular assumptions. Nevertheless, when brought into practice, a perceived “ranking” on the performance is usually perceived by engineers developing machine learning applications. Questions that naturally arise are what kinds of biases the real world has and in which ways can we take advantage from them. Using exploratory data analysis (EDA) on classification examples, we gather insight on some traits that set apart algorithms, datasets and evaluation measures and to what extent the NFL theorem, a theoretical result, applies under typical real-world constraints.     

Using OR to adapt supply chain management best practices to humanitarian logistics

The demand for humanitarian aid is extraordinarily large and it is increasing. In contrast, the funding for humanitarian operations does not seem to be increasing at the same rate. Humanitarian logistics has the challenge of allocating scarce resources to complex operations in an efficient way. After acquiring sufficient contextual knowledge, academics can use operations research (OR) to adapt successful supply chain management best practices to humanitarian logistics. We present two cases of OR applications to field vehicle fleet management in humanitarian operations. Our research shows that by using OR to adapt supply chain best practices to humanitarian logistics, significant improvements can be achieved.     

Training realisation and training demand in the Rioja wine sector in Spain

Training has been identified as an effective tool in the development of human resources, especially in the current context of change. This work aims to analyse the realisation and demand on training courses in the Rioja wine sector in Spain. A representative sample of companies in the sector was interviewed using a questionnaire. Two types of analysis were conducted: univariate and a factorial analysis of variance. These analyses allow using a hypothesis test in which three independent variables are controlled – zone, ownership, and winery size. The results show that the realisation and demand for training are focused on topics related to wine production, and not on marketing. No significant differences were found regarding the realisation and demand for training and the independent variables.     

N‐Benzyl‐4‐((heteroaryl)methyl)benzamides: A New Class of Direct NADH‐Dependent 2‐trans Enoyl–Acyl Carrier Protein Reductase (InhA) Inhibitors with Antitubercular Activity

Isoniazid (INH) remains one of the cornerstones of antitubercular chemotherapy for drug‐sensitive strains of M. tuberculosis bacteria. However, the increasing prevalence of multidrug‐resistant (MDR) and extensively drug‐resistant (XDR) strains containing mutations in the KatG enzyme, which is responsible for the activation of INH into its antitubercular form, have rendered this drug of little or no use in many cases of drug‐resistant tuberculosis. Presented herein is a novel family of antitubercular direct NADH‐dependent 2‐trans enoyl–acyl carrier protein reductase (InhA) inhibitors based on an N‐benzyl‐4‐((heteroaryl)methyl)benzamide template; unlike INH, these do not require prior activation by KatG. Given their direct InhA target engagement, these compounds should be able to circumvent KatG‐related resistance in the clinic. The lead molecules were shown to be potent inhibitors of InhA and showed activity against M. tuberculosis bacteria. This new family of inhibitors was found to be chemically tractable, as exemplified by the facile synthesis of analogues and the establishment of structure–activity relationships. Furthermore, a co‐crystal structure of the initial hit with the enzyme is disclosed, providing valuable information toward the design of new InhA inhibitors for the treatment of MDR/XDR tuberculosis.     

Self-nucleation of isotactic poly(1-butene) in the trigonal modification

Isotactic poly(1-butene) (i-PBu) is a polyolefin of industrial relevance which exhibits an interesting polymorphism. Upon cooling from the relaxed melt at atmospheric pressure, a tetragonal phase (Form II) develops. However, being metastable, this structure slowly evolves upon aging in the stable trigonal modification (Form I). Another trigonal modification, denoted Form I′, can also crystallize directly from the melt if proper conditions are met, e.g., high pressure, low tacticity, ultrathin samples etc. In this work, we aim to verify whether, by acting on the nucleation stage via a proper thermal history, the direct formation of the trigonal polymorph from the melt in a bulk sample is possible. Nucleation of i-PBu has been tailored by means of the self-nucleation technique, imposed on previously aged samples (Form I). DSC and temperature-resolved WAXS show that different crystallization pathways can be observed, depending on the residual concentration of Form I self-nuclei. With decreasing self-nucleation temperature we first encounter “cross-nucleation” of the tetragonal crystal on trigonal nuclei, followed by concomitant crystallization of Form II and Form I′ and eventually the sole formation of the trigonal modification (Form I′).     

Modeling of the dynamic mechanical properties of semicrystalline thermoplastic matrix composites

The modeling of the viscoelastic properties of semicrystalline polymeric matrix composites is discussed. Different mathematical models are applied to describe the development of the storage modulus, measured by dynamic mechanical analysis, as a function of the crystalline content during isothermal crystallization experiments. Best results are obtained with an empirical power law model developed in this work. The application of the Halpin-Tsai equation and of a theoretical model, based on a combination of parallel-series arrangements of viscoelastic elements representing the crystalline and amorphous phases, is also discussed. The main objective of this research is the comparison between experimental results obtained by differential scanning calorimetry and dynamic mechanical analysis during crystallization processes.     

Amine-impregnated natural zeolite as filler in mixed matrix membranes for CO2/CH4 separation

Flat mixed matrix membranes (MMMs) comprising polysulfone and clinoptilolite-type natural zeolite were prepared by casting. Zeolite was modified with three alkylamines: ethanolamine (EA), bis(2-hydroxypropyl)amine (BHPA), and polyethylenimine (PEI) by the impregnation method. Impregnated zeolite samples were characterized by X-ray diffraction, Fourier transform infrared spectroscopy, and N₂ adsorption–desorption. The alkylamine loading extent determined by thermogravimetric analysis was 5.2, 4.8, and 8.5% for EA, BHPA, and PEI, respectively. Analyses of MMMs showed that the incorporation of impregnated zeolite affected the glass-transition temperature (T_g) and mixed-gas transport properties. In this regard, a decreasing trend of the T_g values from 185.5 °C for the polymeric membrane up to 176.6 °C for Clino-EA-based MMM was recorded. In addition, the gas separation performance was evaluated at two different feed pressures. At 50 psi, MMMs showed an enhancement up to 30% on the CO₂ permeability (22.79 Barrer) and 55% on the CO₂/CH₄ selectivity (45.78) in comparison with the polymeric membrane (CO₂ permeability 17.34 Barrer; CO₂/CH₄ selectivity 29.38). These values varied depending on the alkylamine, BHPA being the most selective. © 2019 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2020 , 137, 48286.     

Discriminating power of the hydrocarbon content from virgin olive oil of extremadura cultivars

Samples of virgin olive oils (105) from seven Extremaduran olive varieties (Cacereña, Carrasqueña, Cornezuelo, Corniche, Morisca, Picual, and Verdial de Badajoz) in three stage of maturity (green, semi-ripe, and ripe) were collected and the alkane, alkene, and sesquiterpene contents determined. There were significant differences at the 0.01 probability level in most of the hydrocarbons, both by variety and by state of maturity. Discriminant analysis applied to 70 samples explained 72.3% of the variance between the different groups of varieties and allowed 90% of the samples to be classified according to their variety. The acceptability of the model was verified against the remaining 35 samples, giving a mean level of correct classification of 94%.