No description available.

The goal of this thesis is to study and model the Diel Vertical Migration (DVM) pattern using machine learning methods. We choose an Almost Periodic Function as the mathematical model and fit the monthly averaged migration data into a 5-term Fourier series whose coefficients and frequency are functions of time. The resulting function captures the general characteristics of the DVM pattern whose period is similar yet undergoes gradual changes over time. Further correlation analyses show that the monthly averaged distribution of zooplankton and various environmental factors are strongly correlated. Therefore, we adjust the function so that the coefficients and frequency are functions of environmental factors. Besides, we also examine the pattern on finer time scales using classification algorithms. We build classifiers which predict zooplankton existence at different depths based on a set of environmental measurements. Experiments demonstrate that both of the above methods are valid in modeling the DVM pattern.

A frequent problem in physics and engineering is that of determining a mathematical model for the dynamic performance of a system. It is particularly useful to be able to make measurements which enable such a model to follow changes in the system dynamics in the course of normal operation. Linear control theory, although now being replaced by a more general approach, can still form the basis for such system analysis. Cross correlating signals from a linear process can give a great deal of information about the process dynamics without injecting any test disturbances, or, when test signals are possible, cross correlation can be used to recover dynamic information in the presence of considerable background noise. The use of specially constructed test signals can make cross correlation a powerful technique in the identification of dynamic systems.

Although e-enterprise has been a main stream, there are many stories about the resistance of employees. On the other hand, by the times of knowledge economy age and the popularity of e-learning, a lot of people join the online classes or traditional classes with the support of e-learning system after their office hours. Actually, e-learning system is one kind of e-enterprise systems. If the e-learning experience can let the students be more knowledgeable to the e-enterprise, it will be very helpful to proceed e-enterprise project. The purpose of this research is to probe whether the experience of e-learning would impact the cognition of e-enterprises development. This research adopted quantitative method to explore the relationship between e-learning experience and e-enterprise cognition. A questionnaire survey has been sent to the students who have used the e-learning systems in online courses or traditional courses. According to statistic analyses, the major conclusions are as follows: 1. The e-learning experiences have significant impact on e-enterprise cognition. 2. The impact of the experience of using discussion tools of e-learning systems on e-enterprise cognition will not be interfered by student¡¦s studying attitude. 3. However, The impact of the experience of using supplementary tools of e-learning systems on e-enterprise cognition will be interfered by student¡¦s studying attitude.

The El Nino event causes the changes of the ocean and atmosphere system that induces the rainfall unusual increasing or reduction in some areas and then cause local lives and economical losses. Previous studies have found that the El Nino actually applies impact on the rainfall, however most of the studies focus on the impact of separated stations but little on regional variation. The study on the other hand focus on the variation of the rainfall based on the climatic classification primarily and the physiographic region position auxiliary during the El Nino event and La Nina events. The main method of this research is the correlation analysis, when the correlation coefficient draws close to +1, it mean that the rainfall is positive relative with the parameter of the El Nino, and when the correlation coefficient draws close to -1, it mean that the rainfall is negativity relative with the parameter of the El Nino event. The analysis parameters of the El Nino event index include the sea water temperature and anomaly of every area in Pacific Ocean, sea water surface temperature difference of two areas opposite, Southern Oscillation index and Multivariate ENSO Index. It is found in the study that the best parameter of the El Nino event is the sea water temperature difference of (Nino1¡Ï2- Nino34). The result showed the most climatic classifications have good relation with the parameter of the El Nino, especially winter-dry climatic classifications is the best. Because the result of the research influence on the season variation, it is not to conclude the relation with the El Nino event. It is need to study deeply for calculating the rainfall of the areas where influenced by the El Nino event.

Determining protein structure and function experimentally is both costly and time consuming. Transferring function-related protein annotations based on homology-based methods is relatively straightforward for proteins that have sequence identity of more than 40%. However, there are many proteins in the "twilight zone" where sequence similarity with any other protein is very weak, while being structurally similar to several. Such cases require methods that are capable of using and exploiting both sequence and structural similarity. To understand ways of how such methods can and should be designed is the focus of this study. In this thesis, models that use both sequence and structure features are applied on two protein prediction problems that are particularly challenging when relying on sequence alone. Enzyme classification benefits from both kinds of features because on one hand, enzymes can have identical function with limited sequence similarity while on the other hand, proteins with similar fold may have disparate enzyme class annotation. This thesis shows that the full integration of protein sequence and structure-related features (via the use of kernels) automatically places proteins with similar biological properties closer together, leading to superior classification accuracy using Support Vector Machines. Disulfide-bonds link residues in a protein structure, but may appear distant in sequence. Sequence similarity reflecting such structural properties is thus very hard to detect. It is sufficient for the structure to be similar for accurate prediction of disulfide-bonds, but such information is very scarce and predictors that rely on protein structure are not nearly as useful as those operating on sequence alone. This thesis proposes a novel approach based on Kernel Canonical Correlation Analysis that uses structural features during training only. It does so by finding sequence representations that correlate with structural features that are essential for a disulfide bond. The resulting representations enable high prediction accuracy for a range of disulfide-bond problems. The proposed model thus taps the advantage of structural features without requiring protein structure to be available in the prediction process. The merits of this approach should apply to a number of open protein structure prediction problems.

This simple quantitative descriptive case controlled research compared cases (subjects at risk for acute confusion) with controls (subjects without the attribute); comparison was made on the exposure to potential contributing factors suspected of causing acute confusion, for example, heavy smoking, or the number of alcoholic drinks consumed per day. Case-control studies were also retrospective, because they focused on conditions in the past that might have caused subjects to become cases, rather than controls. The basic purpose of this research design was essentially the same as that of experimental research: to determine the relationships among variables. This report demonstrates that, with relatively good adherence by the nursing team, proactive screening using a structured risk assessment protocol can be successfully implemented for medical patients. This assessment was associated with a statistically significant 50 per cent reduction in the incidence of acute confusion in the intervention group, compared with usual care retrospectively. Reduction in acute confusion was not associated with shortened length of stay, but length of stay was often predetermined by protocol or critical pathway. Correlation analysis demonstrated that risk screening appeared most effective in preventing or reducing acute confusion in patients without preadmission dementia or ADL impairment. In patients with significant preadmission impairment, the stress of hospitalisation may be sufficient to precipitate an episode, despite otherwise optimal management. Less-impaired patients may require additional insults to precipitate acute confusion, some of which are avertable by risk screening and subsequent early intervention. Determined risk indicators were consistent throughout the four year timeframe set for this research project. This demonstrated that although there were multiple patient types presenting to this clinical area, they were consistently the same over a longitudinal timeframe. It meant they were reproducible, which gave this research additional strength. Also, based on the descriptive statistics, this research has shown that in this clinical area where intervention was introduced the combination did have a positive impact on annual numbers of acute confusion. In summary, these findings suggest that without risk screening and the direction for appropriate management the likelihood of an episode can more than double. In the three subgroups expected to pose the greatest challenges for the risk assessment (i.e. those 70 years or older, those with suspected drug dependency, and those with symptomatic infection), risk assessment retained excellent sensitivity, (a) (d) specificity, and relevant correlation with reduction of episodes. This research has demonstrated throughout that high risk screening and associated intervention based on the risk indicator can decrease the annual number of actual episodes of acute confusion. Interventions to prevent or reduce an episode of acute confusion, as outlined by Wakefield (2002) and this research, definitely increases as a result of high risk screening. Beyond doubt, from both the literature reviewed and the findings of this research, is that risk screening does need to be adapted to the individual clinical setting and cannot be generic.

No description available.

Master's thesis "Segmental evaluation of standing posture using accelerometers" is focused on assessment of relationships between body segments during quiet standing with varied sensory afferentation according to CTSIB test. Theoretical part reviews prevailing theoretical work regarding posture, stability, equilibrium and balance and discusses the possibilities of posture assessment with accent on accelerometry and jerk based metrics. The objective of experimental part is to discover influence of sensory afferentation on following aspects: acceleration and jerk of body segments (head, thorax, sacrum and shins); linear correlation of acceleration between body segments; linear correlation of jerk between body segments; jerk magnitude ratio of individual segments. The results demonstrate increasing jerk of sacrum and shins with more demanding postural condition. Jerk linear correlations vary from 0,05 to 0,3 which contradicts the simple inverted pendulum hypothesis of body movement during quiet standing. Higher correlations are registered in the standing on the foam surface compared to firm surface. This relation is accentuated while standing with closed eyes. Higher correlations suggest increasing tendency to the single segment strategy. All above mentioned data are statistically significant. Jerk of...

Machine analysis of human facial and body language is a challenging topic in computer vision, impacting on important applications such as human-computer interaction and visual surveillance. In this thesis, we present research building towards computational frameworks capable of automatically understanding facial expression and behavioural body language. The thesis work commences with a thorough examination in issues surrounding facial representation based on Local Binary Patterns (LBP). Extensive experiments with different machine learning techniques demonstrate that LBP features are efficient and effective for person-independent facial expression recognition, even in low-resolution settings. We then present and evaluate a conditional mutual information based algorithm to efficiently learn the most discriminative LBP features, and show the best recognition performance is obtained by using SVM classifiers with the selected LBP features. However, the recognition is performed on static images without exploiting temporal behaviors of facial expression. Subsequently we present a method to capture and represent temporal dynamics of facial expression by discovering the underlying low-dimensional manifold. Locality Preserving Projections (LPP) is exploited to learn the expression manifold in the LBP based appearance feature space. By deriving a universal discriminant expression subspace using a supervised LPP, we can effectively align manifolds of different subjects on a generalised expression manifold. Different linear subspace methods are comprehensively evaluated in expression subspace learning. We formulate and evaluate a Bayesian framework for dynamic facial expression recognition employing the derived manifold representation. However, the manifold representation only addresses temporal correlations of the whole face image, does not consider spatial-temporal correlations among different facial regions. We then employ Canonical Correlation Analysis (CCA) to capture correlations among face parts. To overcome the inherent limitations of classical CCA for image data, we introduce and formalise a novel Matrix-based CCA (MCCA), which can better measure correlations in 2D image data. We show this technique can provide superior performance in regression and recognition tasks, whilst requiring significantly fewer canonical factors. All the above work focuses on facial expressions. However, the face is usually perceived not as an isolated object but as an integrated part of the whole body, and the visual channel combining facial and bodily expressions is most informative. Finally we investigate two understudied problems in body language analysis, gait-based gender discrimination and affective body gesture recognition. To effectively combine face and body cues, CCA is adopted to establish the relationship between the two modalities, and derive a semantic joint feature space for the feature-level fusion. Experiments on large data sets demonstrate that our multimodal systems achieve the superior performance in gender discrimination and affective state analysis.