Ensemble of Efficient Minimal Learning Machines for Classification and Regression

Minimal Learning Machine (MLM) is a recently proposed supervised learning algorithm with performance comparable to most state-of-the-art machine learning methods. In this work, we propose ensemble methods for classification and regression using MLMs. The goal of ensemble strategies is to produce more robust and accurate models when compared to a single classifier or regression model. Despite its successful application, MLM employs a computationally intensive optimization problem as part of its test procedure (out-of-sample data estimation). This becomes even more noticeable in the context of ensemble learning, where multiple models are used. Aiming to provide fast alternatives to the standard MLM, we also propose the Nearest Neighbor Minimal Learning Machine and the Cubic Equation Minimal Learning Machine to cope with classification and single-output regression problems, respectively. The experimental assessment conducted on real-world datasets reports that ensemble of fast MLMs perform comparably or superiorly to reference machine learning algorithms.     

Microstructure features affecting mechanical properties and corrosion behavior of a hypoeutectic Al–Ni alloy

The aim of this article is to analyze the influence of microstructural parameters on the mechanical properties and corrosion behavior of a hypoeutectic Al–Ni alloy. Experimental results include secondary dendrite arm spacing, corrosion potential, current density, pitting potential, ultimate tensile strength and yield strength. It was found that cooling rates during solidification of about 0.6 °C/s and 8 °C/s can provide secondary dendritic spacings of 7 μm and 16 μm, respectively. Although the microstructure having their phases finely and homogeneously distributed was shown to induce better mechanical properties and higher pitting potential, its general corrosion resistance decreased when compared with the corresponding results of the coarser microstructure.     

Microstructural evolution during upward and downward transient directional solidification of hypomonotectic and monotectic Al–Bi alloys

Upward directional unsteady-state solidification experiments were performed with both a hypomonotectic Al–2.0 wt% Bi alloy and a monotectic Al–3.2 wt% Bi alloy. Besides, the monotectic composition (3.2 wt% Bi) was directionally solidified under downward transient heat flow conditions, which enables the effects of melt convection on the final microstructure to be evaluated since the collective downward movement of Bi-rich particles is favored in such case. This is due to the density differences between the two coexisting liquid phases. The thermal parameters such as cooling rate, growth rate and thermal gradient were experimentally determined by data collected from cooling curves recorded along the casting length. The monotectic features observed in the Al–3.2 wt% Bi alloy castings, i.e. the interphase spacing and Bi-rich particles diameter were correlated with the growth rate and thermal gradient. The cell spacing was experimentally determined for the Al–2.0 wt% Bi alloy as a function of both the cooling rate and tip growth rate. These experimental data were compared with the main predictive cellular growth models from the literature. A comparison between upward and downward unsteady-state solidification results for the interphase spacing and Bi-rich particles diameter has also been conducted.     

Electrochemical corrosion characterization of Al-Ni alloys in a dilute sodium chloride solution

The aim of this study is to characterize the electrochemical corrosion resistance of Al-Ni alloy samples which were directionally solidified under upward unsteady state heat flow conditions. Electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization techniques were used to analyze the corrosion resistance in a dilute 0.05 M NaCl solution at 25 °C. Equivalent circuit analyses were also conducted. It was found that microstructural features such as the dendritic arrangement and the distribution of Al₃Ni intermetallic particles have important roles on both the resulting pitting potential and the general corrosion resistance of Al-Ni alloys.     

Corrosion resistance of directionally solidified Al–6Cu–1Si and Al–8Cu–3Si alloys castings

The aim of this article is to compare the electrochemical corrosion resistance of two as-cast Al–6 wt.% Cu–1 wt.% Si and Al–8 wt.% Cu–3 wt.% Si alloys considering both the solutes macrosegregation profiles and the scale of the microstructure dendritic arrays. A water-cooled unidirectional solidification system was used to obtain the as-cast samples. Electrochemical impedance spectroscopy (EIS) and potentiodynamic anodic polarization techniques were used to analyze the corrosion resistance in a 0.5 M NaCl solution at 25 °C. It was found that the Al–8Cu–3Si alloy has better electrochemical corrosion resistance than the Al–6Cu–1Si alloy for any position along the casting length. At the castings regions where the Cu inverse profile prevailed (up to about 10 mm from the castings surface) the corrosion current density decreased up to 2.5 times with the decrease in the secondary dendrite arm spacing.     

The effect of cooling rate on the dendritic spacing and morphology of Ag3Sn intermetallic particles of a SnAg solder alloy

The size and morphology of intermetallic compounds of Sn–Ag solder alloys can have a significant influence on the mechanical strength of solder joints. The aim of the present study is to characterize the as-cast microstructure of a Sn–2 wt.% Ag solder alloy, and to correlate the resulting scale of the dendritic matrix and the morphology of the Ag₃Sn intermetallic compound (IMC) with the corresponding solidification cooling rate. Pre-heated low-carbon steel molds and a water-cooled solidification apparatus were used permitting a significant range of solidification cooling rates to be experimentally examined. It is shown that under very slow cooling conditions (0.02 °C/s) the microstructure of the sample is formed by a coarse dendritic matrix and a mixture of fiber and plate-like Ag₃Sn IMC in the interdendritic region with the fibers located along the board line separating the matrix. For cooling rates from 0.15 to 1.15 °C/s a mixture of spheroid and fiber-like IMC and secondary dendrite arm spacings between 15 and 40 μm, with the spheroids located in the center of the interdendritic region. At higher cooling rates, of about 8 °C/s only Ag₃Sn spheroids (of about 0.5 μm in diameter) prevail in the eutectic mixture.     

Correlating surface roughness and vibration on plunge cylindrical grinding of steel

Since the wear of a grinding wheel has a direct effect on the workpiece vibration and both have effect on the workpiece quality, the main goal of this work is to study the relation between the process vibration signals and the workpiece quality (mean roughness, circularity and burning) as the grinding wheel gets worn, in an attempt to use these signals to decide the exact moment to dress the wheel. In order to reach this goal, several experiments were carried out in a plunge cylindrical grinding operation of an AISI 52100 quenched and tempered steel, having as input variables the dressing overlap ratio, the spark out time and the workpiece velocity. The output variables were the workpiece surface roughness and circularity and also the process vibration during both, the cutting phase and the spark out phase of the grinding cycle. The main conclusions were: (1) it is possible to have good workpiece quality even with a vibration level much higher than that obtained with a recently dressed wheel; (2) vibration during cutting phase and at the end of complete spark out can be used to monitor the wheel condition at least when high dressing overlap ratio is used; and (3) the decrease in the spark out time makes the vibration at the end of spark out increase a lot, but does not cause such a damage in surface roughness. This fact makes the use of partial spark out feasible in some situations.     

Wetting Behavior and Mechanical Properties of Sn-Zn and Sn-Pb Solder Alloys

A comparative experimental study of the main features of the eutectic Sn-Pb alloy and Sn-Zn alloys was carried out with a view to application of the latter alloys as alternative solder materials. The resulting microstructures, mechanical properties (ultimate tensile strength and elongation), and wettability behavior (spreading area and contact angle) of a hypoeutectic Sn-Zn (Sn-4wt.%Zn), a hypereutectic Sn-Zn (Sn-12wt.%Zn), and the eutectic Sn-9wt.%Zn alloy were examined and compared with the corresponding results of the conventional Sn-40wt.%Pb solder alloy. It was found that, of the Sn-Zn alloys examined, the eutectic Sn-9wt.%Zn alloy offers a compromise between lower wettability and higher mechanical strength.     

Solidification microstructure-dependent hydrogen generation behavior of Al–Sn and Al–Fe alloys in alkaline medium

This work deals with the development of quantitative correlations of hydrogen evolution performance with solidification microstructural and thermal parameters in Al–1Sn, Al–2Sn, Al–1Fe, and Al-1.5Fe [wt.%] alloys. The cellular growth as a function of growth and cooling rates is evaluated using power type experimental laws, which allow determining representative intervals of microstructure length scale for comparison purposes with the results of immersion tests in 5 wt%NaOH solution. For both Al alloys systems, hydrogen evolution becomes slower as the alloy solute content increased. However, for a given alloy composition, whereas a more homogeneous distribution of Sn-rich particles promotes faster hydrogen generation using Al–Sn alloys, coarsening of Al₆Fe IMCs (intermetallic compounds) fibers favors hydrogen production using Al–Fe alloys. When solidification conditions that result in a range of cellular spacings within 16 and 19 μm are considered, the specific hydrogen production of the Al-1wt.%Fe alloy is higher than that of the other studied alloys.     

Contributions from cognitive engineering to requirements specifications for complex sociotechnical systems: A case study in the context of healthcare in Brazil

This paper aims at presenting a case study on the use of human factors and ergonomics to enhance requirement specifications for complex sociotechnical system support tools through enhancing the understanding of human performance within the business domain and the indication of high‐value requirements candidates to information technology support. This work uses methods based on cognitive engineering to build representations of the business domain, highlighting workers’ needs, and contributing to the improvement of software requirements specifications, used in the healthcare domain. As the human factors discipline fits between human sciences and technology design, we believe that its concepts can be combined with software engineering to improve understanding of how people work, enabling the design of better information technology.