Coal-oil mixture from a synthetic oil obtained by hydroliquefaction of a South African coal

A coal-oil mixture (COM) was prepared from a South African coal and a synthetic oil derived from the same coal by hydroliquefaction. The settling and rheological properties of this mixture were compared with those of a mixture of the same coal with fuel oil. The synthetic oil mixture showed greater settling stability, comparable with that of a commercial COM containing additives. The increased stability may be related to polar compounds present in the oil and on the coal surface.     

Thermal stability of Al/AlGaAs and Al/GaAs/AlGaAs (MBE) Schottky barriers

Outstanding stability has been observed in Al/Al/sub x/Ga/sub 1-x/As and Al/GaAs/Al/sub x/Ga/sub 1-x/As (x=0.25) Schottky barriers prepared by depositing Al in situ by MBE on annealing up to 400 degrees C. Conventionally evaporated barriers have been fabricated and compared with epitaxial ones. The changes in barrier height and ideality factor induced by annealing are reported.<>     

Stable redesign of predictive control

The paper deals with I/O versions of receding horizon controllers based on the minimization of multistep quadratic costs with the constraint that the terminal state goes to zero. The resulting control law yields stable closed-loop systems under sharp conditions. Simulation results are presented to both verify the theoretical analysis and relate the new control law with GPC     

Primary lymphoma of the breast. Three cases of non-Hodgkin lymphoma

Primary and sole breast lymphoma is a very rare disease. With the review of our series of 616 cases operated on for breast cancer, only 3 cases (0.48%) of primary breast non-Hodgkin lymphoma (LNH) have been observed. The authors outline the problems concerning diagnosis and therapy of this rare disease: pathological and immunohistochemical aspects are discussed. They stress the importance of staging in order to plan a correct multidisciplinary approach.     

Cervical electrostimulation in some vestibular diseases

Vestibular disorders may be treated by means of drugs, surgical tools or rehabilitation. Cervical electrostimulation may be regarded as physical therapy and its activity is connected to neural pathways between cervical receptors and vestibular nuclei. In present study electrodes were placed lateral to cervical column on the opposite side of vestibular deficit as pointed by electronystagmography; stimulus duration was 18 msec and frequency 100 Hz. Every stimulation lasts for 30 minutes and were repeated twice a week for a total number of 10. The ability to adjust vestibulo-ocular reflex was evaluated by means of a Compensatory Index. The following materials were enrolled in the study: 20 normal subjects were observed before and after a cycle of stimulations: 24 patients suffering from acute labyrinthitis or neuritis, 48 cervical vertigo, 8 cervical traumatic injures and 3 patients operated for acoustical neuroma. All these groups were coupled with control groups composed of the same pathology and the same number of subjects. Statistical treatment of reports was evaluated by t-student test before therapy after one month and after 3 months. Results point out good effects of electrostimulation on vestibulo-ocular reflex compensation, especially in cervical vertigo and labyrinthitis or neuritis with respect to control groups. No important results were obtained in neuroma group and in normal subjects. Electrostimulation may act on vestibular nuclei by direct spino-vestibular pathway but also by repeated decompensations mainly mediated via the cerebellum or spino-reticular connections.     

Enlarging the Margins in Perceptron Decision Trees

Capacity control in perceptron decision trees is typically performed by controlling their size. We prove that other quantities can be as relevant to reduce their flexibility and combat overfitting. In particular, we provide an upper bound on the generalization error which depends both on the size of the tree and on the margin of the decision nodes. So enlarging the margin in perceptron decision trees will reduce the upper bound on generalization error. Based on this analysis, we introduce three new algorithms, which can induce large margin perceptron decision trees. To assess the effect of the large margin bias, OC1 (Journal of Artificial Intelligence Research, 1994, 2, 1–32.) of Murthy, Kasif and Salzberg, a well-known system for inducing perceptron decision trees, is used as the baseline algorithm. An extensive experimental study on real world data showed that all three new algorithms perform better or at least not significantly worse than OC1 on almost every dataset with only one exception. OC1 performed worse than the best margin-based method on every dataset.     

Pheomelanin Effect on UVB Radiation-Induced Oxidation/Nitration of l-Tyrosine

Pheomelanin is a natural yellow-reddish sulfur-containing pigment derived from tyrosinase-catalyzed oxidation of tyrosine in presence of cysteine. Generally, the formation of melanin pigments is a protective response against the damaging effects of UV radiation in skin. However, pheomelanin, like other photosensitizing substances, can trigger, following exposure to UV radiation, photochemical reactions capable of modifying and damaging cellular components. The photoproperties of this natural pigment have been studied by analyzing pheomelanin effect on oxidation/nitration of tyrosine induced by UVB radiation at different pH values and in presence of iron ions. Photoproperties of pheomelanin can be modulated by various experimental conditions, ranging from the photoprotection to the triggering of potentially damaging photochemical reactions. The study of the photomodification of l-Tyrosine in the presence of the natural pigment pheomelanin has a special relevance, since this tyrosine oxidation/nitration pathway can potentially occur in vivo in tissues exposed to sunlight and play a role in the mechanisms of tissue damage induced by UV radiation.     

Latent Semantic Kernels

Kernel methods like support vector machines have successfully been used for text categorization. A standard choice of kernel function has been the inner product between the vector-space representation of two documents, in analogy with classical information retrieval (IR) approaches.Latent semantic indexing (LSI) has been successfully used for IR purposes as a technique for capturing semantic relations between terms and inserting them into the similarity measure between two documents. One of its main drawbacks, in IR, is its computational cost.In this paper we describe how the LSI approach can be implemented in a kernel-defined feature space.We provide experimental results demonstrating that the approach can significantly improve performance, and that it does not impair it.