Selecting the forgetting factor in subset autoregressive modelling

Abstract. Conventional methods to determine the forgetting factors in autoregressive (AR) models are mostly based on arbitrary or personal choices. In this paper, we present two procedures which can be used to select the forgetting factor in subset AR modelling. The first procedure uses the bootstrap to determine the value of a fixed forgetting factor. The second procedure starts from this base and applies the time-recursive maximum likelihood estimation to a variable forgetting factor. In one illustration using real exchange rates, we demonstrate the effect of the forgetting factor in subset AR modelling on ex ante forecasting of non-stationary time series. In a second illustration, these two procedures are applied to time-update forecasts for a stock market index. Subset AR models not including a forgetting factor act as a set of benchmarks for assessing ex ante forecasting performance, and consistently improved forecasting performance is demonstrated for these proposed procedures.     

Catalytic conversion of methane to benzene over Mo/ZSM-5

Dehydroaromatization of methane to benzene occurs over a 2 wt% Mo/ZSM-5 catalyst at 700C under non-oxidizing conditions. Following an initial induction period, during which CH₄ reactant reduces the original Mo⁶⁺ ions in the zeolite to Mo₂C and deposition of coke occurs, a benzene selectivity of 70% at a CH₄ conversion of 8–10% could be sustained for more than 16 h. X-ray photoelectron spectroscopy and X-ray powder diffraction measurements indicate that the reduced Mo is highly dispersed in the channels of the zeolite. Initial activation of CH₄ reactant occurs on Mo₂C sites, leading to the formation of C₂H₄ as the primary product. The latter then undergoes subsequent oligomerization reactions on acidic sites of the zeolite to form aromatic products.     

Shear rates investigation in a scraped surface heat exchanger

The electrodiffusion technique was performed in order to investigate the shear rate on a scraped surface heat exchanger. Microelectrodes were placed inside: the walls of the outer cylinder; the inlet and outlet bowls; the rotor and the blades. Highly viscous Newtonian fluid (Emkarox HV45 solutions) and non-Newtonian model fluid (aqueous solutions of CMC) were used. The electrodiffusion method allowed us to measure wall shear rates. Maximum shear rate was observed at the scraping surface and caused by blades scraping, high shear rate was also measured on the leading edge of the blades. In the other parts of the exchanger, shear rate remained low but the development of Taylor vortices completely modified the scraped surface heat exchangers behaviour inside the surface of the bowls. A dimensionless representation of the friction factor was established for the inner and outer wall surface of the exchanger.     

Solvent fractionation of petroleum pitch for mesophase formation

Petroleum pitch (Ashland A240) has been subjected to thermal treatment and solvent fractionation to produce refined pitches to be evaluated as impregnants for C-C composites. The solvent fractions were obtained by sequential Soxhlet extraction with solvents such as hexane, benzene, and pyridine. Thermal treatment to 408 °C produced a mesophase pitch (≈50% mesophase); an appreciable portion of the mesophase was soluble in strong solvents. There were substantial differences in chemical composition and in pyrolysis behaviour of the fractions. As the depth of fraction increased, the pyrolysis yield and bloating increased, and the microstructure of the coke became finer until glassy microconstituents were formed in the deepest fractions.     

Electrocatalyst materials for fuel cells based on the polyoxometalates—K7 or H7[(P2W17O61)Fe(H2O)] and Na12 or H12[(P2W15O56)2Fe4(H2O)2]

Free acids of the iron substituted heteropoly acids (HPA), H₇[(P₂W₁₇O₆₁)Fe^III(H₂O)] (HFe1) and H₁₈[(P₂W₁₅O₅₆)₂Fe^III₂(H₂O)₂] (HFe2) were prepared from the salts K₇[(P₂W₁₇O₆₁)Fe^III(H₂O)] (KFe1) and Na₁₂[(P₂W₁₅O₅₆)₂Fe^III₄(H₂O)₂] (NaFe4), respectively. The iron-substituted HPA were adsorbed on to XC-72 carbon based GDLs to form HPA doped GDEs after water washing with HPA loadings of ca. 1 μmol. The HPA was detected throughout the GDL by EDX. Solution electrochemistry of the free acids are reported for the first time in sulfate buffer, pH 1-3. The hydrogen oxidation reaction was catalyzed by KFe1 at 0.33 V, with an exchange current density of 38 mA/cm². Moderate activity for the oxygen reduction reaction was observed for the iron substituted HPA, which was dramatically improved by selectively removing oxygen atoms from the HPA by cycling the fuel cell cathode under N₂ followed by reoxidation to give a restructured oxide catalyst. The nanostructured oxide achieved an OCV of 0.7 V with a Tafel slope of 115 mV/decade. Cycling the same catalysts in oxygen resulted in an improved catalyst/ionomer/carbon configuration with a slightly higher Tafel slope, 128 mV/decade but a respectable current density of 100 mA/cm² at 0.2 V.     

Study of the Interfacial Reaction of Styrene/Maleic Anhydride Copolymerand (Amine)-Terminated Butadiene-Acrylonitrile Copolymer Using FTIR Imaging Techniques

The interfacial reaction between styrene/maleic anhydride copolymer (SMA) and amine-terminated butadiene-acrylonitrile copolymer (ATBA) was observed using Fourier transform infrared (FTIR) imaging spectroscopy. The anhydride and amine reacted to form an imide. Each component was detected using a characteristic wavenumber, which was 1601 cm -1 for SMA, 2237 cm -1 for ATBA, and 1701 cm -1 for the imide. FTIR images were taken as the reaction proceeded at 150, 160, 170, and 180°C. At low temperatures (150 and 160°C), diffusion-controlled kinetics were observed since interdiffusion between the reactants did not appear in the images. On the other hand, both the diffusion front and the reaction front are observed in the images at high reaction temperatures (170 and 180°C), which indicates that the kinetics became reaction-controlled rather than diffusion-controlled. Absorbance profiles were extracted from the images and used for the calculation of the extent of reaction. The data were analyzed using the Frederickson and Milner theory based on the assumption of diffusion-controlled kinetics, and this theory fit the low reaction temperature data.     

To what end? Drip irrigation and the water–energy–food nexus in Morocco

This article draws on three case studies of drip irrigation adoption in Morocco to consider the water–energy–food nexus concept from a bottom-up perspective. Findings indicate that small farmers' adoption of drip irrigation is conditional, that water and energy efficiency does not necessarily reduce overall consumption, and that adoption of drip irrigation (and policies supporting it) can create winners and losers. The article concludes that, although the water–energy–food WEF nexus concept may offer useful insights, its use in policy formulation should be tempered with caution. Technical options that appear beneficial at the conceptual level can have unintended consequences in practice, and policies focused on issues of scarcity and efficiency may exacerbate other dimensions of poverty and inequality.     

Capturing crossings: Convex hulls of segment and plane intersections

We give a simple O(nlogn) algorithm to compute the convex hull of the (possibly Θ(n²)) intersection points in an arrangement of n line segments in the plane. We also show an arrangement of dn hyperplanes in d-dimensions whose arrangement has Θ(nd−1) intersection points on the convex hull.