Fitting spatial models in the R package: hglm

We present a new version of the hglm package for fittinghierarchical generalized linear models (HGLM) with spatially correlated random effects. A CAR family for conditional autoregressive random effects was implemented. Eigen decomposition of the matrix describing the spatial structure (e.g. the neighborhood matrix) was used to transform the CAR random effectsinto an independent, but heteroscedastic, gaussian random effect. A linear predictor is fitted for the random effect variance to estimate the parameters in the CAR model.This gives a computationally efficient algorithm for moderately sized problems (e.g. n<5000).

A study of slag corrosion of oxides and oxide-carbon refractories during steel refining

The use of ceramic material as refractories in the manufacturing industry is a common practice worldwide. During usage, for example in the production of steel, these materials do experience severe working conditions including high temperatures, low pressures and corrosive environments. This results in lowered service lives and high consumptions of these materials. This, in turn, affects the productivity of the whole steel plant and thereby the cost. In order to investigate how the service life can be improved, studies have been carried out for refractories used in the inner lining of the steel ladles. More specifically, from the slag zone, where the corrosion is most severe. By combining thermodynamic simulations, plant trails and post-mortem studies of the refractories after service, vital information about the behaviour of the slagline refractories during steel refining and the causes of the accelerated wear in this ladle area has been achieved. The results from these studies show that the wear of the slagline refractories of the ladle is initiated at the preheating station, through reduction-oxidation reactions. The degree of the decarburization process is mostly dependent on the preheating fuel or the environment. For refractories without antioxidants, refractory decarburization is slower when coal gas is used in ladle preheating than when a mixture of oil and air is used. In addition, ladle preheating of the refractories without antioxidants leads to direct wear of the slagline refractories. This is due to the total loss of the matrix strength, which results in a sand-like product. Thermal chemical changes that take place in the slagline refractories are due to the MgO-C reaction as well as the formation of liquid phases from impurity oxides. In addition, the decrease in the system pressure during steel refining makes the MgO-C reaction take place at the steel refining temperatures. This reduces the refractory’s resistance to corrosion. This is a serious problem for both the magnesia-carbon and dolomite-carbon refractories. The studies of the reactions between the slagline refractories and the different slag compositions showed that slags rich in iron oxide lead mostly to the oxidation of carbon/graphite in the carbon-containing refractories. This leads to an increased porosity and wettability and therefore an enhanced penetration of slag into the refractory structure. If the slag contains high contents of alumina and or silica (such as the steel refining slag), reactions between the slag components and the dolomite-carbon refractory are promoted. This leads to the formation of low-temperature melting phases such as calcium-aluminates and silicates. The state of these reaction products during steel refining leads to an accelerated wear of the dolomite-carbon refractory. The main products of the reactions between the magnesia-carbon refractory and the steel refining slag are MgAl2O4 spinels, and calcium-aluminates, and silicates. Due to the good refractory properties of MgAl2O4 spinels, the slag corrosion resistance of the magnesiacarbon refractory is promoted.

An approach to evaluate product verifications

Companies implement a module product assortment as a part of their strategy to, among others, shorten lead-times, increase the product quality and to create more product variants with fever parts. However, the increased number of variants becomes a challenging task for the personnel responsible for the product verifications. By implementing verifications at module level, so called MPV (Module Property Verification) several advantages ensue. The advantages is not only a decrease in cost of verifications, but also a decrease in repair times, occupied space, storages with spare parts, and repair tools. Further, MPV also give an increased product quality due to an increased understanding of which defects that may occur. As an approach to implement MPV, this paper discusses defects and verification processes based on a study at a Swedish company. It also describes a matrix which is used to map relations between company specific cost drivers and so called verification factors. The matrix may indicate cost drivers which have a large impact on the total cost of product verifications.

Movies, Mind and Meaning : Studying Audience and Favourite Films

Construction of identity and meaning is becoming increasingly important in both media studies and religion scholarship. (Lövheim, 2004) Meaning construction outside traditional religion has become more interesting for religious studies and what individuals in the audience do with all messages circulated through media in everyday life has attended increasing interest within media studies (Stout and Buddenbaum, 2001). Motion pictures, soap operas and advertising are all examples of media contents which generate ideas among its audience which to a various degree are used as resources within the construction of identity (Jansson, 2001). The investigation of what modern humankind’s world views look like and what components they are composed of, in this context seems to be an important topic of investigation (Holm and Björkqvist, 1996). The ways in which the development of media has effected the daily lives of individuals is interest as is the nature of the self and the ways in which the process of self-formation is affected by the profusion of mediated materials (Thompson, 1995). Film and religion are my interest within this larger frame. The topic is not exactly new but the combination of film and religion has during the last ten years resulted in a rapidly growing number of books by scholars interested in this field (Lyden, 2003). One growing focus is on the role that films can and do play within the emerging and developing valuesystem of people in the West today (Marsh, 2004). The British theologian Clive Marsh’s point of departure is very similar to my own. Viewers bring to a film life-experience, immediate concerns and worldviews and the exploration of this interplay between movies and the interpreting process of meaning making is the very focus in this paper. Theoretically, the semeiological model of Alf Linderman is combined with cultural cognitive approaches used by a number of Scandinavian media scholars developing perspectives in audience theory (Linderman, 1996, Höijer and Werner, 1998). 13 individuals, their favourite movie and what it means to them in their life My aim is to examine how individuals comprehend film and what the meaning process look like. In this paper I present the outcome of 13 interviews with young people about their favourite film. I suggest how it is possible to interpret how they interrelate film comprehension with their personal beliefs and their culturally constructed worldview from a sociocognitive point of view. Examples of films chosen range from Disneys Lion King (1994), sciencefiction and fantasy successes like The Matrix (1999) and Lord of the Rings (2001) or the next best movie ever according to www.IMdb.com The Shawshank Redemption (1994) as well as the Swedish blockbuster Så som i himmelen (2004), aka “As in Heaven”.

Fitting conditional and simultaneous autoregressive spatial models in hglm

We present a new version (> 2.0) of the hglm package for fitting hierarchical generalized linear models (HGLMs) with spatially correlated random effects. CAR() and SAR() families for conditional and simultaneous autoregressive random effects were implemented. Eigen decomposition of the matrix describing the spatial structure (e.g., the neighborhood matrix) was used to transform the CAR/SAR random effects into an independent, but eteroscedastic, Gaussian random effect. A linear predictor is fitted for the random effect variance to estimate the parameters in the CAR and SAR models. This gives a computationally efficient algorithm for moderately sized problems.