Avaliação estatística de métodos de poda aplicados em neurônios intermediários da rede neural MLP

There are several papers on pruning methods in the artificial neural networks area. However, with rare exceptions, none of them presents an appropriate statistical evaluation of such methods. In this article, we proved statistically the ability of some methods to reduce the number of neurons of the hidden layer of a multilayer perceptron neural network (MLP), and to maintain the same landing of classification error of the initial net. They are evaluated seven pruning methods. The experimental investigation was accomplished on five groups of generated data and in two groups of real data. Three variables were accompanied in the study: apparent classification error rate in the test group (REA); number of hidden neurons, obtained after the application of the pruning method; and number of training/retraining epochs, to evaluate the computational effort. The non-parametric Friedman's test was used to do the statistical analysis.

Análise da incerteza na representação de classes de cobertura do solo urbano resultantes da aplicação de uma rede neural artificial

The great diversity of materials that characterizes the urban environment determines a structure of mixed classes in a classification of multiespectral images. In that sense, it is important to define an appropriate classification system using a non parametric classifier, that allows incorporating non spectral (such as texture) data to the process. They also allow analyzing the uncertainty associated to each class from the output alues of the network calculated in relation to each class. Considering these properties, an experiment was carried out. This experiment consisted in the application of an Artificial Neural Network aiming at the classification of the urban land cover of Presidente Prudente and the analysis of the uncertainty in the representation of the mapped thematic classes. The results showed that it is possible to discriminate the variations in the urban land cover through the application of an Artificial Neural Network. It was also possible to visualize the spatial variation of the uncertainty in the attribution of classes of urban land cover from the generated representations. The class characterized by a defined pattern as intermediary related to the impermeability of the urban soil presented larger ambiguity degree and, therefore, larger mixture.

Bone mineralization in Brazilian adolescents: The years of maximum bone mass incorporation

Puberty is the fundamental period for bone mass (BM) acquisition. In this period mineralization is found to increase with levels of high bone formation. The critical years of intense bone anabolism deserve special attention, as adequate gain could minimize fracture risk in later years. The objective of this work was to study bone mineral content (BMC) and bone mineral density (BMD) in male adolescents with age bracket and maturation level. Sixty-one healthy male 10 to 19 year-olds were evaluated for calcium intake, weight, stature, BMI, puberty stage and BMC and BMD in the lumbar spine and femur. BM was measured by bone densitometry (DXA). Calcium intake was calculated by recording 3 days diet. Puberty stage was defined as per Tanner. Descriptive statistics was used with means and standard deviations, linear correlation, and analysis of variance for comparison between age groups, and the Tukey test (p<0.05). Linear correlation was positive and indicated body weight as the main correlation variable with BMD in both studied locations (p<0.01). BMC and BMD increased with age, differences were significant from 14 to 15 years, and when adolescents reached Tanner stage G4. These results showed a pronounced increase in bone mineralization, with the years after 14 to 15 being critical for BM acquisition in Brazilian adolescents.

Role of annexin 1 gene expression in mouse craniofacial bone development

BACKGROUND: Annexin 1 is a 37-kDa protein that has complex intra- and extracellular effects. To discover whether the absence of this protein alters bone development, we monitored this event in the annexin-A1 null mice in comparison with littermate wild-type controls. METHODS: Radiographic and densitometry methods were used for the assessment of bone in annexin-A1 null mice at a gross level. We used whole-skeleton staining, histological analysis, and Western blotting techniques to monitor changes at the tissue and cellular levels. RESULTS: There were no gross differences in the appendicular skeleton between the genotypes, but an anomalous development of the skull was observed in the annexin-A1 null mice. This was characterized in the newborn annexin-A1 null animals by a delayed intramembranous ossification of the skull, incomplete fusion of the interfrontal suture and palatine bone, and the presence of an abnormal suture structure. The annexin-A1 gene was shown to be active in osteocytes during this phase and COX-2 was abundantly expressed in cartilage and bone taken from annexin-A1 null mice. CONCLUSIONS: Expression of the annexin-A1 gene is important for the normal development of the skull in mice, possibly through the regulation of osteoblast differentiation and a secondary effect on the expression of components of the cPLA2-COX-2 system. © 2007 Wiley-Liss, Inc.

Urban sprawl and energy use for transportation in the largest Brazilian cities

A study of the relationships between the amount of energy consumed for transportation purposes and a few selected variables related to urban form and socioeconomic characteristics of some of the largest Brazilian cities is conducted in this work. The studied cities include all 27 state capitals regardless of their size and population and 184 urban areas each with more than 20,000 inhabitants located in the state of São Paulo. Two different techniques were applied for data analyses: a more traditional regression analysis approach and artificial neural networks. In general, the results found in the analyses conducted here support the assumption that urban sprawl increases the energy use for transportation. In the case of the 27 state capitals, the analysis indicated that two spatial variables have a strong impact on the energy consumed for urban transportation: urban density and the ratio between the longest distances in the east-west and north-south directions. In the case of the 184 urbanized areas we also reached a similar conclusion. In that case, however, income and employment level apparently have a stronger influence on the amount of energy consumed. The results of the present study stress the importance of physical planning in developing country cities in order to reduce energy use for transportation. © 2007 International Energy Initiative, Inc.

Os efeitos de um programa de atividade física de carga progressiva nas propriedades físicas e na força óssea de ratas osteopênicas

Background: Many studies have shown that physical exercises are able to stimulate bone formation and increase bone mass, constituting a therapeutic modality to treat bone loss due to osteoporosis. However, some points about the intensity, duration and frequency of the exercises remain confusing and contradictory. Thus, the aim of this study was to determine the effects of a progressive loading exercise program on femur of osteopenic rats. To induce osteopenia we used the animal model of ovariectomy (OVX). Forty animals was studied and divided into 4 groups: sham-operated sedentary (SS); ovariectomy-sedentary (OS); sham-operated training (ST) and ovariectomy training (OT). The trained groups performed jumps into water: 4 series of 10 jumps each, with an overload of 50% to 80% of the animal's body weight, during 8 weeks. Femora were submitted to a physical properties evaluation, a biomechanical test, calcium and phosphorus content measurement and a morphometric histological evaluation. Results: osteopenic animals showed a decrease of bone strength and lower values of bone weights, bone density and calcium content. The exercised osteopenic rats showed higher values of geometrical, physical properties, bone strength and calcium content compared to controls. The results of the present study indicate that the progressive loading exercise program had stimulatory effects on femora of osteopenic rats. It seems that the intensity and duration of the protocol used produced bone structural adaptations, which contributed to reverse bone loss due to ovariectomy.

Neural network approach for surface roughness prediction in surface grinding

Several systems are currently tested in order to obtain a feasible and safe method for automation and control of grinding process. This work aims to predict the surface roughness of the parts of SAE 1020 steel ground in a surface grinding machine. Acoustic emission and electrical power signals were acquired by a commercial data acquisition system. The former from a fixed sensor placed near the workpiece and the latter from the electric induction motor that drives the grinding wheel. Both signals were digitally processed through known statistics, which with the depth of cut composed three data sets implemented to the artificial neural networks. The neural network through its mathematical logical system interpreted the signals and successful predicted the workpiece roughness. The results from the neural networks were compared to the roughness values taken from the worpieces, showing high efficiency and applicability on monitoring and controlling the grinding process. Also, a comparison among the three data sets was carried out.

Radial basis function networks with quantized parameters

A RBFN implemented with quantized parameters is proposed and the relative or limited approximation property is presented. Simulation results for sinusoidal function approximation with various quantization levels are shown. The results indicate that the network presents good approximation capability even with severe quantization. The parameter quantization decreases the memory size and circuit complexity required to store the network parameters leading to compact mixed-signal circuits proper for low-power applications. ©2008 IEEE.

Environment mapping for mobile robots navigation using hierarchical neural network and omnivision

Autonomous robots must be able to learn and maintain models of their environments. In this context, the present work considers techniques for the classification and extraction of features from images in joined with artificial neural networks in order to use them in the system of mapping and localization of the mobile robot of Laboratory of Automation and Evolutive Computer (LACE). To do this, the robot uses a sensorial system composed for ultrasound sensors and a catadioptric vision system formed by a camera and a conical mirror. The mapping system is composed by three modules. Two of them will be presented in this paper: the classifier and the characterizer module. The first module uses a hierarchical neural network to do the classification; the second uses techiniques of extraction of attributes of images and recognition of invariant patterns extracted from the places images set. The neural network of the classifier module is structured in two layers, reason and intuition, and is trained to classify each place explored for the robot amongst four predefine classes. The final result of the exploration is the construction of a topological map of the explored environment. Results gotten through the simulation of the both modules of the mapping system will be presented in this paper. © 2008 IEEE.

RBF circuits based on folded cascode differential pairs

We propose new circuits for the implementation of Radial Basis Functions such as Gaussian and Gaussian-like functions. These RBFs are obtained by the subtraction of two differential pair output currents in a folded cascode configuration. We also propose a multidimensional version based on the unidimensional circuits. SPICE simulation results indicate good functionality. These circuits are intended to be applied in the implementation of radial basis function networks. One possible application of these networks is transducer signal conditioning in aircraft and spacecraft vehicles onboard telemetry systems. Copyright 2008 ACM.

Effects of Long-Term FK 506 Therapy on the Alveolar Bone and Cementum of Rats

Cyclosporine (CsA) and tacrolimus (FK 506) exert complex, incompletely understood actions on bone. The objective of the study was to evaluate the effects of long-term tacrolimus therapy on the periodontium. Rats were treated for 60, 120, 180, and 240 days with daily subcutaneous injections of 1 mg/kg body weight of FK 506. After the experimental period, we obtained serum levels of calcium and alkaline phosphatase (ALP). After histological processing, the alveolar bone and cementum, as well as volume densities of bone (Vb) and osteoclasts (Vo), were assessed at the regions of the lower first molar. There was a tendency toward a statistically significant decrease in ALP levels with FK 506; however, serum calcium levels increased during the long periods. At 60, 180, and 240 days of treatment with FK 506, we did not observe Vb and Vo alterations. At 120 days of treatment, there was an evident decrease in Vb, but it did not show alveolar bone loss. We did not observe any alterations of cementum among rats treated with FK 506. It may be concluded that FK 506 administration did not induce side effects on the periodontium. © 2009 Elsevier Inc. All rights reserved.

A neural approach to evaluate the effect of lightning in power transformers

This paper proposes the application of computational intelligence techniques to assist complex problems concerning lightning in transformers. In order to estimate the currents related to lightning in a transformer, a neural tool is presented. ATP has generated the training vectors. The input variables used in Artificial Neural Networks (ANN) were the wave front time, the wave tail time, the voltage variation rate and the output variable is the maximum current in the secondary of the transformer. These parameters can define the behavior and severity of lightning. Based on these concepts and from the results obtained, it can be verified that the overvoltages at the secondary of transformer are also affected by the discharge waveform in a similar way to the primary side. By using the tool developed, the high voltage process in the distribution transformers can be mapped and estimated with more precision aiding the transformer project process, minimizing empirics and evaluation errors, and contributing to minimize the failure rate of transformers. © 2009 The Berkeley Electronic Press. All rights reserved.

Immunohistochemical, tomographic and histological study on onlay bone graft remodeling. Part II: Calvarial bone

Objectives: Little information is available on the molecular events that occur during graft incorporation over time. The calvarial bone (Cb) grafts have been reported to produce greater responses compared with other donor regions in maxillofacial reconstructions, but the scientific evidences for this are still lacking. The objectives of this study are (1) to study the morphological pattern of Cb onlay bone grafts and compare them with the biological events through immunohistochemical responses and (2) to establish the effects of perforations in maintaining the volume and bone density of the receptor bed. Material and methods: Sixty New Zealand White rabbits were submitted to Cb onlay bone grafts on the mandible. In 30 rabbits, the receptor bed was perforated (perforated group), while for the remaining animals the bed was kept intact (non-perforated group). Six animals from each group were sacrificed at 5, 7, 10, 20 and 60 days after surgery. Histological sections from the grafted area were prepared for immunohistochemical and histological analyses. Immuno-labeling was found for proteins Osteoprotegerin (OPG), receptor activator of nuclear factor-κβ ligand (RANKL), alkaline phosphatase (ALP), osteopontin (OPN), vascular endothelial growth factor (VEGF), tartrate-resistant acid phosphatase (TRAP), Type I collagen (COL I) and osteocalcin (OC). The tomography examination [computerized tomography (CT) scan] was conducted just after surgery and at the sacrifice. Results: The histological findings revealed that the perforations contributed to higher bone deposition during the initial stages at the graft-receptor bed interface, accelerating the graft incorporation process. The results of the CT scan showed lower resorption for the perforated group (P≤0.05), and both groups showed high bone density rates at 60 days. This set of evidences is corroborated by the immunohistochemical outcomes indicating that proteins associated with revascularization and osteogenesis (VEGF, OPN, TRAP and ALP) were found in higher levels in the perforated group. Conclusions: These findings indicate that the bone volume of calvarial grafts is better maintained when the receptor bed is perforated, probably resulting from more effective graft revascularization and greater bone deposition. The process of bone resorption peaked between 20 and 60 days post-operatively in both groups although significantly less in the perforated group. © 2009 John Wiley & Sons A/S.

Harmonic content identification based on neural method for single phase power systems

An alternative method is presented in this paper to identify the harmonic components of non-linear loads in single phase power systems based on artificial neural networks. The components are identified by analyzing the single phase current waveform in time domain in half-cycle of the ac voltage source. The proposed method is compared to the fast Fourier transform. Simulation and experimental results are presented to validate the proposed approach.

Advances in sickle cell disease treatment: From drug discovery until the patient monitoring

Sickle Cell Disease (SCD) is one of the most prevalent hematological diseases in the world. Despite the immense progress in molecular knowledge about SCD in last years few therapeutical sources are currently available. Nowadays the treatment is performed mainly with drugs such as hydroxyurea or other fetal hemoglobin inducers and chelating agents. This review summarizes current knowledge about the treatment and the advancements in drug design in order to discover more effective and safe drugs. Patient monitoring methods in SCD are also discussed. © 2011 Bentham Science Publishers Ltd.