Application of Neural Network Methods in the Prediction of Stock Markets

The purpose of the research is to define practical profit which can be achieved using neural network methods as a prediction instrument. The thesis investigates the ability of neural networks to forecast future events. This capability is checked on the example of price prediction during intraday trading on stock market. The executed experiments show predictions of average 1, 2, 5 and 10 minutes’ prices based on data of one day and made by two different types of forecasting systems. These systems are based on the recurrent neural networks and back propagation neural nets. The precision of the predictions is controlled by the absolute error and the error of market direction. The economical effectiveness is estimated by a special trading system. In conclusion, the best structures of neural nets are tested with data of 31 days’ interval. The best results of the average percent of profit from one transaction (buying + selling) are 0.06668654, 0.188299453, 0.349854787 and 0.453178626, they were achieved for prediction periods 1, 2, 5 and 10 minutes. The investigation can be interesting for the investors who have access to a fast information channel with a possibility of every-minute data refreshment.

Regression-based techniques for statistical decision making in single-case designs

The present study evaluates the performance of four methods for estimating regression coefficients used to make statistical decisions regarding intervention effectiveness in single-case designs. Ordinary least squares estimation is compared to two correction techniques dealing with general trend and one eliminating autocorrelation whenever it is present. Type I error rates and statistical power are studied for experimental conditions defined by the presence or absence of treatment effect (change in level or in slope), general trend, and serial dependence. The results show that empirical Type I error rates do not approximate the nominal ones in presence of autocorrelation or general trend when ordinary and generalized least squares are applied. The techniques controlling trend show lower false alarm rates, but prove to be insufficiently sensitive to existing treatment effects. Consequently, the use of the statistical significance of the regression coefficients for detecting treatment effects is not recommended for short data series.

Evaluation of Software Development Practices through Error Analysis

This thesis studies evaluation of software development practices through an error analysis. The work presents software development process, software testing, software errors, error classification and software process improvement methods. The practical part of the work presents results from the error analysis of one software process. It also gives improvement ideas for the project. It was noticed that the classification of the error data was inadequate in the project. Because of this it was impossible to use the error data effectively. With the error analysis we were able to show that there were deficiencies in design and analyzing phases, implementation phase and in testing phase. The work gives ideas for improving error classification and for software development practices.

Performance and Methodological Aspects in Positron Emission Tomography

Performance standards for Positron emission tomography (PET) were developed to be able to compare systems from different generations and manufacturers. This resulted in the NEMA methodology in North America and the IEC in Europe. In practices, the NEMA NU 2- 2001 is the method of choice today. These standardized methods allow assessment of the physical performance of new commercial dedicated PET/CT tomographs. The point spread in image formation is one of the factors that blur the image. The phenomenon is often called the partial volume effect. Several methods for correcting for partial volume are under research but no real agreement exists on how to solve it. The influence of the effect varies in different clinical settings and it is likely that new methods are needed to solve this problem. Most of the clinical PET work is done in the field of oncology. The whole body PET combined with a CT is the standard investigation today in oncology. Despite the progress in PET imaging technique visualization, especially quantification of small lesions is a challenge. In addition to partial volume, the movement of the object is a significant source of error. The main causes of movement are respiratory and cardiac motions. Most of the new commercial scanners are in addition to cardiac gating, also capable of respiratory gating and this technique has been used in patients with cancer of the thoracic region and patients being studied for the planning of radiation therapy. For routine cardiac applications such as assessment of viability and perfusion only cardiac gating has been used. However, the new targets such as plaque or molecular imaging of new therapies require better control of the cardiac motion also caused by respiratory motion. To overcome these problems in cardiac work, a dual gating approach has been proposed. In this study we investigated the physical performance of a new whole body PET/CT scanner with NEMA standard, compared methods for partial volume correction in PET studies of the brain and developed and tested a new robust method for dual cardiac-respiratory gated PET with phantom, animal and human data. Results from performance measurements showed the feasibility of the new scanner design in 2D and 3D whole body studies. Partial volume was corrected, but there is no best method among those tested as the correction also depends on the radiotracer and its distribution. New methods need to be developed for proper correction. The dual gating algorithm generated is shown to handle dual-gated data, preserving quantification and clearly eliminating the majority of contraction and respiration movement

On Fault Tolerance Methods for Networks-on-Chip

Technology scaling has proceeded into dimensions in which the reliability of manufactured devices is becoming endangered. The reliability decrease is a consequence of physical limitations, relative increase of variations, and decreasing noise margins, among others. A promising solution for bringing the reliability of circuits back to a desired level is the use of design methods which introduce tolerance against possible faults in an integrated circuit. This thesis studies and presents fault tolerance methods for network-onchip (NoC) which is a design paradigm targeted for very large systems-onchip. In a NoC resources, such as processors and memories, are connected to a communication network; comparable to the Internet. Fault tolerance in such a system can be achieved at many abstraction levels. The thesis studies the origin of faults in modern technologies and explains the classification to transient, intermittent and permanent faults. A survey of fault tolerance methods is presented to demonstrate the diversity of available methods. Networks-on-chip are approached by exploring their main design choices: the selection of a topology, routing protocol, and flow control method. Fault tolerance methods for NoCs are studied at different layers of the OSI reference model. The data link layer provides a reliable communication link over a physical channel. Error control coding is an efficient fault tolerance method especially against transient faults at this abstraction level. Error control coding methods suitable for on-chip communication are studied and their implementations presented. Error control coding loses its effectiveness in the presence of intermittent and permanent faults. Therefore, other solutions against them are presented. The introduction of spare wires and split transmissions are shown to provide good tolerance against intermittent and permanent errors and their combination to error control coding is illustrated. At the network layer positioned above the data link layer, fault tolerance can be achieved with the design of fault tolerant network topologies and routing algorithms. Both of these approaches are presented in the thesis together with realizations in the both categories. The thesis concludes that an optimal fault tolerance solution contains carefully co-designed elements from different abstraction levels

Comparison of USEPA 3050B and ISO 14869-1: 2001 digestion methods for sediment analysis by using FAAS and ICP-OES quantification techniques

A study of the partial USEPA 3050B and total ISO 14869-1:2001 digestion methods of sediments was performed. USEPA 3050B was recommended as the simpler method with less operational risk. However, the extraction ability of the method should be taken in account for the best environmental interpretation of the results. FAAS was used to quantify metal concentrations in sediment solutions. The alternative use of ICP-OES quantification should be conditioned by a previous detailed investigation and eventual correction of the matrix effect. For the first time, the EID method was employed for the detection and correction of the matrix effect in sediment ICP-OES analysis. Finally, some considerations were made about the level of metal contamination in the area under study.

Method of numerical correction of errors occasioned by delay of records during the monitoring of environmental variables of interest for animal production

Thermal and air conditions inside animal facilities change during the day due to the influence of the external environment. For statistical and geostatistical analyses to be representative, a large number of points spatially distributed in the facility area must be monitored. This work suggests that the time variation of environmental variables of interest for animal production, monitored within animal facility, can be modeled accurately from discrete-time records. The aim of this study was to develop a numerical method to correct the temporal variations of these environmental variables, transforming the data so that such observations are independent of the time spent during the measurement. The proposed method approached values recorded with time delays to those expected at the exact moment of interest, if the data were measured simultaneously at the moment at all points distributed spatially. The correction model for numerical environmental variables was validated for environmental air temperature parameter, and the values corrected by the method did not differ by Tukey's test at 5% significance of real values recorded by data loggers.

Comparison measures of maps generated by geostatistical methods

This study uses several measures derived from the error matrix for comparing two thematic maps generated with the same sample set. The reference map was generated with all the sample elements and the map set as the model was generated without the two points detected as influential by the analysis of local influence diagnostics. The data analyzed refer to the wheat productivity in an agricultural area of 13.55 ha considering a sampling grid of 50 x 50 m comprising 50 georeferenced sample elements. The comparison measures derived from the error matrix indicated that despite some similarity on the maps, they are different. The difference between the estimated production by the reference map and the actual production was of 350 kilograms. The same difference calculated with the mode map was of 50 kilograms, indicating that the study of influential points is of fundamental importance to obtain a more reliable estimative and use of measures obtained from the error matrix is a good option to make comparisons between thematic maps.

Hérnia diafragmática traumática crônica e fístula gastropleurocutânea

Traumatic diaphragmatic hernia is defined as a laceration of the diaphragm with an abdominal viscera herniation into the thorax. It is usually asymptomatic, with the exception of the cases with obstruction, strangulation, necrosis or perforation of the herniaded viscera. It is classified as acute, latent or chronic, in accordance with the evolutive period. At the latent phase, symptoms are indefinite and the radiological signals, which are suggestive of thoracic affections, are frequent and can induce a diagnosis error, leading to inadequate treatment.This article presents a case of chronic traumatic diaphragmatic hernia which was complicated by a gastricpleuralcutaneous fistula, due to an inadequate thoracic drainage. Considering that this is a chronic affection with an unquestionable surgical indication, due to the complications risk, it is essential to have a detailed diagnostic investigation, which aims at both avoiding an intempestive or inadequate therapeutics behaviour and reducing the affection morbimortality. Recently, the videolaparoscopic approach has proved to be more precise when compared to the other diagnostic methods, by direct visualization of the diaphragmatic laceration, allowing its correction by an immediate suture.

Simulation methods in the modelling of bioaffinity assays

Computational model-based simulation methods were developed for the modelling of bioaffinity assays. Bioaffinity-based methods are widely used to quantify a biological substance in biological research, development and in routine clinical in vitro diagnostics. Bioaffinity assays are based on the high affinity and structural specificity between the binding biomolecules. The simulation methods developed are based on the mechanistic assay model, which relies on the chemical reaction kinetics and describes the forming of a bound component as a function of time from the initial binding interaction. The simulation methods were focused on studying the behaviour and the reliability of bioaffinity assay and the possibilities the modelling methods of binding reaction kinetics provide, such as predicting assay results even before the binding reaction has reached equilibrium. For example, a rapid quantitative result from a clinical bioaffinity assay sample can be very significant, e.g. even the smallest elevation of a heart muscle marker reveals a cardiac injury. The simulation methods were used to identify critical error factors in rapid bioaffinity assays. A new kinetic calibration method was developed to calibrate a measurement system by kinetic measurement data utilizing only one standard concentration. A nodebased method was developed to model multi-component binding reactions, which have been a challenge to traditional numerical methods. The node-method was also used to model protein adsorption as an example of nonspecific binding of biomolecules. These methods have been compared with the experimental data from practice and can be utilized in in vitro diagnostics, drug discovery and in medical imaging.

Night admission is an independent risk factor for mortality in trauma patients - a systemic error approach

ABSTRACTObjective:to assess the impact of the shift inlet trauma patients, who underwent surgery, in-hospital mortality.Methods:a retrospective observational cohort study from November 2011 to March 2012, with data collected through electronic medical records. The following variables were statistically analyzed: age, gender, city of origin, marital status, admission to the risk classification (based on the Manchester Protocol), degree of contamination, time / admission round, admission day and hospital outcome.Results:during the study period, 563 patients injured victims underwent surgery, with a mean age of 35.5 years (± 20.7), 422 (75%) were male, with 276 (49.9%) received in the night shift and 205 (36.4%) on weekends. Patients admitted at night and on weekends had higher mortality [19 (6.9%) vs. 6 (2.2%), p=0.014, and 11 (5.4%) vs. 14 (3.9%), p=0.014, respectively]. In the multivariate analysis, independent predictors of mortality were the night admission (OR 3.15), the red risk classification (OR 4.87), and age (OR 1.17).Conclusion:the admission of night shift and weekend patients was associated with more severe and presented higher mortality rate. Admission to the night shift was an independent factor of surgical mortality in trauma patients, along with the red risk classification and age.

Use of surgical mesh of different compositions in the correction of the abdominal wall defect in rats

Objective: To analyze the performance of two surgical meshes of different compositions during the defect healing process of the abdominal wall of rats. Methods: thirty-three adult Wistar rats were anesthetized and subjected to removal of an area of 1.5 cm x 2 cm of the anterior abdominal wall, except for the skin; 17 animals had the defect corrected by edge-to-edge surgical suture of a mesh made of polypropylene + poliglecaprone (Group U - UltraproTM); 16 animals had the defect corrected with a surgical mesh made of polypropylene + polidioxanone + cellulose (Group P - ProceedTM). Each group was divided into two subgroups, according to the euthanasia moment (seven days or 28 days after the operation). Parameters analyzed were macroscopic (adherence), microscopic (quantification of mature and immature collagen) and tensiometric (maximum tension and maximum rupture strength). Results : there was an increase in collagen type I in the ProceedTM group from seven to 28 days, p = 0.047. Also, there was an increase in the rupture tension on both groups when comparing the two periods. There was a lower rupture tension and tissue deformity with ProceedTM mesh in seven days, becoming equal at day 28. Conclusion : the meshes retain similarities in the final result and more studies with larger numbers of animals must be carried for better assessment.

Comparison of Two- and Three-dimensional Ultrasonography in the Evaluation of Lesion Level in Fetuses with Spina Bifida

Purpose To evaluate the precision of both two- and three-dimensional ultrasonography in determining vertebral lesion level (the first open vertebra) in patients with spina bifida. Methods This was a prospective longitudinal study comprising of fetuses with open spina bifida who were treated in the fetal medicine division of the department of obstetrics of Hospital das Clínicas of the Universidade de São Paulo between 2004 and 2013. Vertebral lesion level was established by using both two- and three-dimensional ultrasonography in 50 fetuses (two examiners in each method). The lesion level in the neonatal period was established by radiological assessment of the spine. All pregnancies were followed in our hospital prenatally, and delivery was scheduled to allow immediate postnatal surgical correction. Results Two-dimensional sonography precisely estimated the spina bifida level in 53% of the cases. The estimate error was within one vertebra in 80% of the cases, in up to two vertebrae in 89%, and in up to three vertebrae in 100%, showing a good interobserver agreement. Three-dimensional ultrasonography precisely estimated the lesion level in 50% of the cases. The estimate error was within one vertebra in 82% of the cases, in up to two vertebrae in 90%, and in up to three vertebrae in 100%, also showing good interobserver agreement. Whenever an estimate error was observed, both two- and three-dimensional ultrasonography scans tended to underestimate the true lesion level (55.3% and 62% of the cases, respectively). Conclusions No relevant difference in diagnostic performance was observed between the two- and three-dimensional ultrasonography. The use of three-dimensional ultrasonography showed no additional benefit in diagnosing the lesion level in the fetuses with spina bifida. Errors in both methods showed a tendency to underestimate lesion level.

Statistical analysis of survey-based event history data with application to modeling of unemployment duration

Longitudinal surveys are increasingly used to collect event history data on person-specific processes such as transitions between labour market states. Surveybased event history data pose a number of challenges for statistical analysis. These challenges include survey errors due to sampling, non-response, attrition and measurement. This study deals with non-response, attrition and measurement errors in event history data and the bias caused by them in event history analysis. The study also discusses some choices faced by a researcher using longitudinal survey data for event history analysis and demonstrates their effects. These choices include, whether a design-based or a model-based approach is taken, which subset of data to use and, if a design-based approach is taken, which weights to use. The study takes advantage of the possibility to use combined longitudinal survey register data. The Finnish subset of European Community Household Panel (FI ECHP) survey for waves 1–5 were linked at person-level with longitudinal register data. Unemployment spells were used as study variables of interest. Lastly, a simulation study was conducted in order to assess the statistical properties of the Inverse Probability of Censoring Weighting (IPCW) method in a survey data context. The study shows how combined longitudinal survey register data can be used to analyse and compare the non-response and attrition processes, test the missingness mechanism type and estimate the size of bias due to non-response and attrition. In our empirical analysis, initial non-response turned out to be a more important source of bias than attrition. Reported unemployment spells were subject to seam effects, omissions, and, to a lesser extent, overreporting. The use of proxy interviews tended to cause spell omissions. An often-ignored phenomenon classification error in reported spell outcomes, was also found in the data. Neither the Missing At Random (MAR) assumption about non-response and attrition mechanisms, nor the classical assumptions about measurement errors, turned out to be valid. Both measurement errors in spell durations and spell outcomes were found to cause bias in estimates from event history models. Low measurement accuracy affected the estimates of baseline hazard most. The design-based estimates based on data from respondents to all waves of interest and weighted by the last wave weights displayed the largest bias. Using all the available data, including the spells by attriters until the time of attrition, helped to reduce attrition bias. Lastly, the simulation study showed that the IPCW correction to design weights reduces bias due to dependent censoring in design-based Kaplan-Meier and Cox proportional hazard model estimators. The study discusses implications of the results for survey organisations collecting event history data, researchers using surveys for event history analysis, and researchers who develop methods to correct for non-sampling biases in event history data.

Methods of genetic diversity creation and functional display for directed evolution experiments

Protein engineering aims to improve the properties of enzymes and affinity reagents by genetic changes. Typical engineered properties are affinity, specificity, stability, expression, and solubility. Because proteins are complex biomolecules, the effects of specific genetic changes are seldom predictable. Consequently, a popular strategy in protein engineering is to create a library of genetic variants of the target molecule, and render the population in a selection process to sort the variants by the desired property. This technique, called directed evolution, is a central tool for trimming protein-based products used in a wide range of applications from laundry detergents to anti-cancer drugs. New methods are continuously needed to generate larger gene repertoires and compatible selection platforms to shorten the development timeline for new biochemicals. In the first study of this thesis, primer extension mutagenesis was revisited to establish higher quality gene variant libraries in Escherichia coli cells. In the second study, recombination was explored as a method to expand the number of screenable enzyme variants. A selection platform was developed to improve antigen binding fragment (Fab) display on filamentous phages in the third article and, in the fourth study, novel design concepts were tested by two differentially randomized recombinant antibody libraries. Finally, in the last study, the performance of the same antibody repertoire was compared in phage display selections as a genetic fusion to different phage capsid proteins and in different antibody formats, Fab vs. single chain variable fragment (ScFv), in order to find out the most suitable display platform for the library at hand. As a result of the studies, a novel gene library construction method, termed selective rolling circle amplification (sRCA), was developed. The method increases mutagenesis frequency close to 100% in the final library and the number of transformants over 100-fold compared to traditional primer extension mutagenesis. In the second study, Cre/loxP recombination was found to be an appropriate tool to resolve the DNA concatemer resulting from error-prone RCA (epRCA) mutagenesis into monomeric circular DNA units for higher efficiency transformation into E. coli. Library selections against antigens of various size in the fourth study demonstrated that diversity placed closer to the antigen binding site of antibodies supports generation of antibodies against haptens and peptides, whereas diversity at more peripheral locations is better suited for targeting proteins. The conclusion from a comparison of the display formats was that truncated capsid protein three (p3Δ) of filamentous phage was superior to the full-length p3 and protein nine (p9) in obtaining a high number of uniquely specific clones. Especially for digoxigenin, a difficult hapten target, the antibody repertoire as ScFv-p3Δ provided the clones with the highest affinity for binding. This thesis on the construction, design, and selection of gene variant libraries contributes to the practical know-how in directed evolution and contains useful information for scientists in the field to support their undertakings.