A deep learning approach with Bayesian optimization and ensemble classifiers for detecting denial of service attacks

Detecting malicious behavior is important for preventing security threats in a computer network. Denial of Service (DoS) is among the popular cyber attacks targeted at web sites of high‐profile organizations and can potentially have high economic and time costs. In this paper, several machine learning methods including ensemble models and autoencoder‐based deep learning classifiers are compared and tuned using Bayesian optimization. The autoencoder framework enables to extract new features by mapping the original input to a new space. The methods are trained and tested both for binary and multi‐class classification on Digiturk and Labris datasets, which were introduced recently for detecting various types of DDoS attacks. The best performing methods are found to be ensembles though deep learning classifiers achieved comparable level of accuracy.     

Investigation of Ionospheric Electron Density Change During Two Partial Solar Eclipses and Its Comparison with Predictions of NeQuick 2 and IRI-2016 Models

Wireless Personal Communications - In the present study, the results obtained by incoherent scatter radar (ISR) and empirical models (NeQuick2 and IRI-2016) of the variations in mid-latitude...     

Maximization of total surface area of Pt-SnO<Subscript>2</Subscript>/Al<Subscript>2</Subscript>O<Subscript>3</Subscript> catalyst by the Taguchi method

The maximization of the total surface area of Pt-SnO₂/Al₂O₃ catalyst was studied by using the Taguchi method of experimental design. The catalysts were prepared by sol-gel method. The effects of HNO₃, H₂O and aluminum nitrate concentrations and the stirring rate on the total surface area were studied at three levels of each. L₉ orthogonal array leading nine experiments was used in the experimental design. The parameter levels that give maximum total surface area were determined and experimentally verified. In the range of conditions studied it was found that, medium levels of HNO₃ and H₂O concentration and lower levels of aluminum nitrate concentration and stirring rate maximize the total surface area.     

The effects of expanded perlite aggregate, silica fume and fly ash on the thermal conductivity of lightweight concrete

Thermal conductivity coefficients of concretes made up of mixtures of expanded perlite and pumice aggregates (PA) were measured. To determine the effect of silica fume (SF) and class C fly ash (FA) on the thermal conductivity of lightweight aggregate concrete (LWAC), SF and FA were added as replacement for cement by decreasing the cement weights in the ratios of 10%, 20% and 30% by weight.The highest thermal conductivity of 0.3178 W/mK was observed with the samples containing only PA and plain cement. It decreased with the increase of SF and FA as replacement for cement. The lowest value of thermal conductivity, which is 0.1472 W/mK, was obtained with the samples prepared with expanded perlite aggregate (EPA) replacement of PA and 70% cement+30% FA replacement of cement. Both SF and FA had a decreasing effect on thermal conductivity. EPA (used in place of PA) also induced a decrease of 43.5% in thermal conductivity of concrete.     

The preparation of core–shell Al2O3/Ni composite powders by electroless plating

Core–shell nanostructured Ni-coated Al₂O₃ composite powders were synthesised by using the electroless plating method. The influence of the chemical components and powder concentration in the Ni coating was investigated by scanning electron microscopy, energy dispersive spectroscopy, and X-ray diffraction techniques. The results show that the concentration of the plating components plays an important role in the formation of core–shell Al₂O₃/Ni composite powders. The nickel content in the composite powders could be effectively controlled by adjusting the nickel chloride content and the concentration of NaH₂PO₂·H₂O in the plating solution. The nanostructure of the crystalline Ni coatings was observed to be very attractive for achieving good bonding between ceramic particles and matrices for composite production.     

Electrochemical preparation and characterization of nickel and zinc-modified poly-2-aminothiazole films on mild steel surface and their corrosion inhibition performance

Poly-2-aminothiazole (pAT) was electrochemically synthesized on a mild steel (MS) specimen from 0.3 M aqueous ammonium oxalate solution containing 0.01 M 2-aminothiazole (2-AT) using cyclic voltammetry technique. The synthesized polymer film was then modified by electrodeposition of 100 μg cm⁻² Ni (MS/pAT–Ni) and Zn (MS/pAT–Zn) on top of the polymer surface. The surface morphologies of the polymer films were examined by scanning electron microscopy (SEM) and atomic force microscopy (AFM). The elemental analysis of the surface films was performed by energy dispersive X-ray spectroscopy (EDX). The effectiveness of the coatings in preventing corrosion of MS in 3.5% NaCl solution was assessed using electrochemical techniques. It was found that the obtained coatings were adherent to the steel surface. The pAT film provided a good corrosion protection against the attack of corrosive environment. Moreover, the modification of pAT film by deposition of Ni and Zn on top of the polymer surface significantly enhances the corrosion protection performance of the polymer film by exhibiting an improved barrier effect against the attack of corrosive environment. The surface morphologies and protection ability of the layers were found to be dependent on the type of deposited metal.     

Selective CO oxidation over promoted Au/γ-Al2O3 catalysts in the presence of CO2 and H2O in the feed

In this work, the effect of promoter type (Mg, Mn, Ce, Co, Fe and Ni) on selective CO oxidation performance of Au/γ-Al₂O₃ was studied with the realistic feed stream containing CO₂ and H₂O. The effects of Au loading, promoter loading, reaction temperature and the feed composition were also investigated. It was found that MgO was the best promoter in the presence of CO₂ and H₂O, and 1.25 wt.% Mg was sufficient for promotion. The CO conversion decreased with the addition of CO₂ while the presence of H₂O had some positive effects.     

Comparison of the structural and magnetic properties of submicron barium hexaferrite powders prepared by molten salt and solid state calcination routes

The effect of crystallinity and particle morphology of the submicron barium hexaferrite (BaFe₁₂O₁₉) powders on the magnetic properties was investigated on powders synthesized by solid-state calcination (BHF-c) and molten salt synthesis (BHF-m) methods. Solid-state calcination route was found to yield agglomerated powders with poor crystallinity, whereas molten salt synthesis resulted in well crystallized powders with an anisometric morphology. The saturation magnetization of the BHF-m and BHF-c samples is 59 emu/g, and 56 emu/g at 300 K, and 90 emu/g, and 86 emu/g at 10 K. The temperature dependence of magnetization of the BHF-m is higher and the increase in magnetocrystal anisotropy with decreasing temperature is also steeper than that of the BHF-c due to the higher crystallinity. The magnetocrystalline anisotropy constant, K, calculated from the Stoner–Wohlfarth theory, of the BHF-m and BHF-c powders is 14.24 and 10.14 HA²/kg, respectively. The higher effective anisotropy, K_eff of the BHF-m is also confirmed through ferromagnetic resonance measurements. In conclusion, the higher crystallinity, slightly higher particle size and anisometric morphology of the BHF-m particles translated into higher magnetic properties and magnetocrystalline anisotropy.