First-principle data-driven models for assessment of motor disorders in Parkinson’s disease

Parkinson’s disease (PD) is an increasing neurological disorder in an aging society. The motor and non-motor symptoms of PD advance with the disease progression and occur in varying frequency and duration. In order to affirm the full extent of a patient’s condition, repeated assessments are necessary to adjust medical prescription. In clinical studies, symptoms are assessed using the unified Parkinson’s disease rating scale (UPDRS). On one hand, the subjective rating using UPDRS relies on clinical expertise. On the other hand, it requires the physical presence of patients in clinics which implies high logistical costs. Another limitation of clinical assessment is that the observation in hospital may not accurately represent a patient’s situation at home. For such reasons, the practical frequency of tracking PD symptoms may under-represent the true time scale of PD fluctuations and may result in an overall inaccurate assessment. Current technologies for at-home PD treatment are based on data-driven approaches for which the interpretation and reproduction of results are problematic.  The overall objective of this thesis is to develop and evaluate unobtrusive computer methods for enabling remote monitoring of patients with PD. It investigates first-principle data-driven model based novel signal and image processing techniques for extraction of clinically useful information from audio recordings of speech (in texts read aloud) and video recordings of gait and finger-tapping motor examinations. The aim is to map between PD symptoms severities estimated using novel computer methods and the clinical ratings based on UPDRS part-III (motor examination). A web-based test battery system consisting of self-assessment of symptoms and motor function tests was previously constructed for a touch screen mobile device. A comprehensive speech framework has been developed for this device to analyze text-dependent running speech by: (1) extracting novel signal features that are able to represent PD deficits in each individual component of the speech system, (2) mapping between clinical ratings and feature estimates of speech symptom severity, and (3) classifying between UPDRS part-III severity levels using speech features and statistical machine learning tools. A novel speech processing method called cepstral separation difference showed stronger ability to classify between speech symptom severities as compared to existing features of PD speech. In the case of finger tapping, the recorded videos of rapid finger tapping examination were processed using a novel computer-vision (CV) algorithm that extracts symptom information from video-based tapping signals using motion analysis of the index-finger which incorporates a face detection module for signal calibration. This algorithm was able to discriminate between UPDRS part III severity levels of finger tapping with high classification rates. Further analysis was performed on novel CV based gait features constructed using a standard human model to discriminate between a healthy gait and a Parkinsonian gait. The findings of this study suggest that the symptom severity levels in PD can be discriminated with high accuracies by involving a combination of first-principle (features) and data-driven (classification) approaches. The processing of audio and video recordings on one hand allows remote monitoring of speech, gait and finger-tapping examinations by the clinical staff. On the other hand, the first-principles approach eases the understanding of symptom estimates for clinicians. We have demonstrated that the selected features of speech, gait and finger tapping were able to discriminate between symptom severity levels, as well as, between healthy controls and PD patients with high classification rates. The findings support suitability of these methods to be used as decision support tools in the context of PD assessment.

How does data quality in a network affect heuristic solutions?

To have good data quality with high complexity is often seen to be important. Intuition says that the higher accuracy and complexity the data have the better the analytic solutions becomes if it is possible to handle the increasing computing time. However, for most of the practical computational problems, high complexity data means that computational times become too long or that heuristics used to solve the problem have difficulties to reach good solutions. This is even further stressed when the size of the combinatorial problem increases. Consequently, we often need a simplified data to deal with complex combinatorial problems. In this study we stress the question of how the complexity and accuracy in a network affect the quality of the heuristic solutions for different sizes of the combinatorial problem. We evaluate this question by applying the commonly used p-median model, which is used to find optimal locations in a network of p supply points that serve n demand points. To evaluate this, we vary both the accuracy (the number of nodes) of the network and the size of the combinatorial problem (p). The investigation is conducted by the means of a case study in a region in Sweden with an asymmetrically distributed population (15,000 weighted demand points), Dalecarlia. To locate 5 to 50 supply points we use the national transport administrations official road network (NVDB). The road network consists of 1.5 million nodes. To find the optimal location we start with 500 candidate nodes in the network and increase the number of candidate nodes in steps up to 67,000 (which is aggregated from the 1.5 million nodes). To find the optimal solution we use a simulated annealing algorithm with adaptive tuning of the temperature. The results show that there is a limited improvement in the optimal solutions when the accuracy in the road network increase and the combinatorial problem (low p) is simple. When the combinatorial problem is complex (large p) the improvements of increasing the accuracy in the road network are much larger. The results also show that choice of the best accuracy of the network depends on the complexity of the combinatorial (varying p) problem.

Combining Satellite And Gauge Precipitation Data With Co-Kriging Method For Jakarta Region

Jakarta is vulnerable to flooding mainly caused by prolonged and heavy rainfall and thus a robust hydrological modeling is called for. A good quality of spatial precipitation data is therefore desired so that a good hydrological model could be achieved. Two types of rainfall sources are available: satellite and gauge station observations. At-site rainfall is considered to be a reliable and accurate source of rainfall. However, the limited number of stations makes the spatial interpolation not very much appealing. On the other hand, the gridded rainfall nowadays has high spatial resolution and improved accuracy, but still, relatively less accurate than its counterpart. To achieve a better precipitation data set, the study proposes cokriging method, a blending algorithm, to yield the blended satellite-gauge gridded rainfall at approximately 10-km resolution. The Global Satellite Mapping of Precipitation (GSMaP, 0.1⁰×0.1⁰) and daily rainfall observations from gauge stations are used. The blended product is compared with satellite data by cross-validation method. The newly-yield blended product is then utilized to re-calibrate the hydrological model. Several scenarios are simulated by the hydrological models calibrated by gauge observations alone and blended product. The performance of two calibrated hydrological models is then assessed and compared based on simulated and observed runoff.

Investigating The Sources Of Fresh Water Affecting The Hydrological Balance Of Lakes Enriquillo And Azuei (Hispaniola) – Data Analysis

The Enriquillo and Azuei are saltwater lakes located in a closed water basin in the southwestern region of the island of La Hispaniola, these have been experiencing dramatic changes in total lake-surface area coverage during the period 1980-2012. The size of Lake Enriquillo presented a surface area of approximately 276 km2 in 1984, gradually decreasing to 172 km2 in 1996. The surface area of the lake reached its lowest point in the satellite observation record in 2004, at 165 km2. Then the recent growth of the lake began reaching its 1984 size by 2006. Based on surface area measurement for June and July 2013, Lake Enriquillo has a surface area of ~358 km2. Sumatra sizes at both ends of the record are 116 km2 in 1984 and 134 km2in 2013, an overall 15.8% increase in 30 years. Determining the causes of lake surface area changes is of extreme importance due to its environmental, social, and economic impacts. The overall goal of this study is to quantify the changing water balance in these lakes and their catchment area using satellite and ground observations and a regional atmospheric-hydrologic modeling approach. Data analyses of environmental variables in the region reflect a hydrological unbalance of the lakes due to changing regional hydro-climatic conditions. Historical data show precipitation, land surface temperature and humidity, and sea surface temperature (SST), increasing over region during the past decades. Salinity levels have also been decreasing by more than 30% from previously reported baseline levels. Here we present a summary of the historical data obtained, new sensors deployed in the sourrounding sierras and the lakes, and the integrated modeling exercises. As well as the challenges of gathering, storing, sharing, and analyzing this large volumen of data in a remote location from such a diverse number of sources.

Data mining por meio de análise de redes, no contexto de filtro colaborativo

Tendo como motivação o desenvolvimento de uma representação gráfica de redes com grande número de vértices, útil para aplicações de filtro colaborativo, este trabalho propõe a utilização de superfícies de coesão sobre uma base temática multidimensionalmente escalonada. Para isso, utiliza uma combinação de escalonamento multidimensional clássico e análise de procrustes, em algoritmo iterativo que encaminha soluções parciais, depois combinadas numa solução global. Aplicado a um exemplo de transações de empréstimo de livros pela Biblioteca Karl A. Boedecker, o algoritmo proposto produz saídas interpretáveis e coerentes tematicamente, e apresenta um stress menor que a solução por escalonamento clássico.

Extraction of Building Roof Contours From LiDAR Data Using a Markov-Random-Field-Based Approach

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Fault section estimation in electric power systems using an optimization immune algorithm

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

An algorithm for arranging response surface designs in small blocks

We consider the problem of blocking response surface designs when the block sizes are prespecified to control variation efficiently and the treatment set is chosen independently of the block structure. We show how the loss of information due to blocking is related to scores defined by Mead and present an interchange algorithm based on scores to improve a given blocked design. Examples illustrating the performance of the algorithm are given and some comparisons with other designs are made. (C) 2000 Elsevier B.V. B.V. All rights reserved.

A machine learning approach for genome-wide prediction of morbid and druggable human genes based on systems-level data

Background: The genome-wide identification of both morbid genes, i.e., those genes whose mutations cause hereditary human diseases, and druggable genes, i.e., genes coding for proteins whose modulation by small molecules elicits phenotypic effects, requires experimental approaches that are time-consuming and laborious. Thus, a computational approach which could accurately predict such genes on a genome-wide scale would be invaluable for accelerating the pace of discovery of causal relationships between genes and diseases as well as the determination of druggability of gene products.Results: In this paper we propose a machine learning-based computational approach to predict morbid and druggable genes on a genome-wide scale. For this purpose, we constructed a decision tree-based meta-classifier and trained it on datasets containing, for each morbid and druggable gene, network topological features, tissue expression profile and subcellular localization data as learning attributes. This meta-classifier correctly recovered 65% of known morbid genes with a precision of 66% and correctly recovered 78% of known druggable genes with a precision of 75%. It was than used to assign morbidity and druggability scores to genes not known to be morbid and druggable and we showed a good match between these scores and literature data. Finally, we generated decision trees by training the J48 algorithm on the morbidity and druggability datasets to discover cellular rules for morbidity and druggability and, among the rules, we found that the number of regulating transcription factors and plasma membrane localization are the most important factors to morbidity and druggability, respectively.Conclusions: We were able to demonstrate that network topological features along with tissue expression profile and subcellular localization can reliably predict human morbid and druggable genes on a genome-wide scale. Moreover, by constructing decision trees based on these data, we could discover cellular rules governing morbidity and druggability.

An alternative approach to solve convergence problems in the backpropagation algorithm

The multilayer perceptron network has become one of the most used in the solution of a wide variety of problems. The training process is based on the supervised method where the inputs are presented to the neural network and the output is compared with a desired value. However, the algorithm presents convergence problems when the desired output of the network has small slope in the discrete time samples or the output is a quasi-constant value. The proposal of this paper is presenting an alternative approach to solve this convergence problem with a pre-conditioning method of the desired output data set before the training process and a post-conditioning when the generalization results are obtained. Simulations results are presented in order to validate the proposed approach.

Simple orbit determination using GPS based on a least-squares algorithm employing sequential givens rotations

A low-cost computer procedure to determine the orbit of an artificial satellite by using short arc data from an onboard GPS receiver is proposed. Pseudoranges are used as measurements to estimate the orbit via recursive least squares method. The algorithm applies orthogonal Givens rotations for solving recursive and sequential orbit determination problems. To assess the procedure, it was applied to the TOPEX/POSEIDON satellite for data batches of one orbital period (approximately two hours), and force modelling, due to the full JGM-2 gravity field model, was considered. When compared with the reference Precision Orbit Ephemeris (POE) of JPL/NASA, the results have indicated that precision better than 9 m is easily obtained, even when short batches of data are used. Copyright (c) 2007.

Screening Brazilian C gasoline quality: Application of the SIMCA chemometric method to gas chromatographic data

A total of 2400 samples of commercial Brazilian C gasoline were collected over a 6-month period from different gas stations in the São Paulo state, Brazil, and analysed with respect to 12 physicochemical parameters according to regulation 309 of the Brazilian Government Petroleum, Natural Gas and Biofuels Agency (ANP). The percentages (v/v) of hydrocarbons (olefins, aromatics and saturated) were also determined. Hierarchical cluster analysis (HCA) was employed to select 150 representative samples that exhibited least similarity on the basis of their physicochemical parameters and hydrocarbon compositions. The chromatographic profiles of the selected samples were measured by gas chromatography with flame ionisation detection and analysed using soft independent modelling of class analogy (SIMCA) method in order to create a classification scheme to identify conform gasolines according to ANP 309 regulation. Following the optimisation of the SIMCA algorithm, it was possible to classify correctly 96% of the commercial gasoline samples present in the training set of 100. In order to check the quality of the model, an external group of 50 gasoline samples (the prediction set) were analysed and the developed SIMCA model classified 94% of these correctly. The developed chemometric method is recommended for screening commercial gasoline quality and detection of potential adulteration. (c) 2007 Elsevier B.V. All rights reserved.

Modelo numérico do terreno obtido por diferentes métodos em cartas planialtimétricas

The digital elevation model is important to determine the slope and land use capability, therefore, a proposal of methodology for acquisition of elevation data contemplating an efficient algorithm to generate a slope map was developed. Thus, it was aimed to obtain and evaluate a digital elevation model without the vetorization of the contours on planialtimetric charts. The area for acquisition of elevation data was Sao Manuel, SP. The data were collected by two methods: level contour vetorization and the gathering of elevation points on the level contour with maximum elevation points. The elevation data were analyzed by geostatistical techniques. Inspite of wide difference in the number of collected points between two methods, the variograms were adjusted to the exponential model and showed a range of approximately 1500 m, which does not justify the wide difficulty in vetorization of the planialtimetric charts, once the data points collected in the area were appropriately distributed, they represented rightly the terrain surface.

PMUs optimized allocation using a tabu search algorithm for fault location in electric power distribution system

Phasor Measurement Units (PMUs) optimized allocation allows control, monitoring and accurate operation of electric power distribution systems, improving reliability and service quality. Good quality and considerable results are obtained for transmission systems using fault location techniques based on voltage measurements. Based on these techniques and performing PMUs optimized allocation it is possible to develop an electric power distribution system fault locator, which provides accurate results. The PMUs allocation problem presents combinatorial features related to devices number that can be allocated, and also probably places for allocation. Tabu search algorithm is the proposed technique to carry out PMUs allocation. This technique applied in a 141 buses real-life distribution urban feeder improved significantly the fault location results. © 2004 IEEE.