Vibroarthrograpyhy (VAG) is an innovative, objective and non-invasive technique to obtain diagnostic information concerning the articular cartilage of knee joints. Knee VAG signals can be detected by putting a contact sensor on the surface of the knee joints during the movement such as flexion and extension. Before this research, there are many VAG group studies that contribute in signal processing and database building. The adaptive segmentation method and autoregressive modeling are developed to segment the nonstationary VAG signals. This thesis tries to investigate the accuracy of some database containing root mean square (RMS) value and intraclass distance (ID) feature parameters of physiological patellofemoral crepitus (PPC) signals. This research is first setting up two diagnosis standards for RMS and ID. According to the two standards, all signals are divided into three types: normal, unknown and injured, and those appear both in normal type of RMS and ID parameters are picked out. The same does the injured type. In conclusion, by checking the anamneses of these signals, we can be aware of the numbers of real normal and real injured in normal type and injured type; therefore the accuracy of the database can be derived. Consequently the accuracy of database in this thesis is quite certifiable.

GDP is used to measure the economic state of a country and accurate forecasts of it is therefore important. Using the Economic Tendency Survey we investigate forecasting quarterly GDP growth using the data mining technique Random Forest. Comparisons are made with a benchmark AR(1) and an ad hoc linear model built on the most important variables suggested by the Random Forest. Evaluation by forecasting shows that the Random Forest makes the most accurate forecast supporting the theory that there are benefits to using Random Forests on economic time series.

Neste trabalho, fazemos uma breve introdução à Econofísica e às grandezas estatísticas relevantes para o estudo de um ativo financeiro. Estas grandezas são estudadas detalhadamente para o índice NYSE Composto. Determinamos o tempo de autocorrelação e o espectro de potência, cujos resultados indicam a presença de uma correlação de curto alcance. Através do expoente de Hurst, investigamos o tipo de correlação presente e detectamos a presença de multifractalidade. A volatilidade do índice NYSE mostrou-se análoga a um processo de Wiener. Por outro lado, a função densidade de probabilidade do índice NYSE foi ajustada por uma distribuição de Lévy simétrica com alpha = 1,47. Apresentamos os modelos de variância autoregressiva ARCH e GARCH. Em particular, focalizamos o modelo Markoviano GARCH(1,1). Este modelo tem três parâmetros de controle. Mostramos que, para o índice NYSE, o uso do tempo de autocorrelação na determinação deste conjunto de parâmetros de controle não é a melhor escolha. Resultados muito mais satisfatórios são obtidos se utilizarmos o sexto momento padronizado, uma vez que o ganho no ajuste da função de autocorrelação temporal é muito mais expressivo. A proposta de utilização do sexto momento é robusta e se aplica tanto ao modelo GARCH Gaussiano quanto ao modelo GARCH Exponencial. Desenvolvemos uma técnica de expansão em série para obter o sexto momento padronizado em função dos três parâmetros de controle. Obtivemos uma expressão analítica exata para a curtose do modelo GARCH Exponencial. Ambas as versões Gaussiana e Exponencial apresentam um desempenho equivalente na descrição da função densidade de probabilidade e da função de autocorrelação temporal. Porém, no que tange às leis de escala temporal, medidas através da probabilidade de retorno à origem, o modelo Exponencial tem, clara e inequivocamente, um melhor desempenho que o modelo Gaussiano, pois apresenta um expoente da lei de escala temporal em bom acordo com o expoente do índice NYSE. / In this thesis, we briefly give an introduction to Econophysics and discuss some important statistical quantities used in the study of a financial asset. This quantities are meticulously studied for the NYSE Composite Index. For its time series, we determine the time autocorrelation and the power spectrum, which show the presence of a short range correlation. By means of the Hurst exponent, we investigate the kind of autocorrelation which is present and we detected the presence of multifractality. The volatility of the NYSE Index show a behavior analogous to a Wiener process. On the other hand, the probability density function was adjusted by a symmetric Lévy distribuition with alpha = 1.47. We present the variance autoregressive ARCH and GARCH models. More specifically, we focus on the Markovian GARCH(1,1) model. This model has three control parameters. We show that, for the NYSE Index, the use of the time autocorrelation to determinate the set of control parameters is not the best choice. Instead, results much more reasonable are obtained if the standardized sixth moment is used, as can be seen by the adjust of the time autocorrelation function. The proposal of the sixth moment is robust and applies for both the Gaussian and the Exponential GARCH models. We developed a series expansion technique to get the standardized sixth moment as a function of the three control parameters. We found an exact analytic expression for the kurtosis of the Exponential GARCH model. Both the Gaussian and the Exponential versions exhibit an equivalent performance in the description of the probability density function and the time autocorrelation function. However, with respect to the time scaling laws (measured by the probability of return to the origin) the Exponential model shows, in a clear and unequivocal way, a better performance than the Gaussian model, since it gives a time horizon exponent much more close to the real NYSE exponent.

Neste trabalho, fazemos uma breve introdução à Econofísica e às grandezas estatísticas relevantes para o estudo de um ativo financeiro. Estas grandezas são estudadas detalhadamente para o índice NYSE Composto. Determinamos o tempo de autocorrelação e o espectro de potência, cujos resultados indicam a presença de uma correlação de curto alcance. Através do expoente de Hurst, investigamos o tipo de correlação presente e detectamos a presença de multifractalidade. A volatilidade do índice NYSE mostrou-se análoga a um processo de Wiener. Por outro lado, a função densidade de probabilidade do índice NYSE foi ajustada por uma distribuição de Lévy simétrica com alpha = 1,47. Apresentamos os modelos de variância autoregressiva ARCH e GARCH. Em particular, focalizamos o modelo Markoviano GARCH(1,1). Este modelo tem três parâmetros de controle. Mostramos que, para o índice NYSE, o uso do tempo de autocorrelação na determinação deste conjunto de parâmetros de controle não é a melhor escolha. Resultados muito mais satisfatórios são obtidos se utilizarmos o sexto momento padronizado, uma vez que o ganho no ajuste da função de autocorrelação temporal é muito mais expressivo. A proposta de utilização do sexto momento é robusta e se aplica tanto ao modelo GARCH Gaussiano quanto ao modelo GARCH Exponencial. Desenvolvemos uma técnica de expansão em série para obter o sexto momento padronizado em função dos três parâmetros de controle. Obtivemos uma expressão analítica exata para a curtose do modelo GARCH Exponencial. Ambas as versões Gaussiana e Exponencial apresentam um desempenho equivalente na descrição da função densidade de probabilidade e da função de autocorrelação temporal. Porém, no que tange às leis de escala temporal, medidas através da probabilidade de retorno à origem, o modelo Exponencial tem, clara e inequivocamente, um melhor desempenho que o modelo Gaussiano, pois apresenta um expoente da lei de escala temporal em bom acordo com o expoente do índice NYSE. / In this thesis, we briefly give an introduction to Econophysics and discuss some important statistical quantities used in the study of a financial asset. This quantities are meticulously studied for the NYSE Composite Index. For its time series, we determine the time autocorrelation and the power spectrum, which show the presence of a short range correlation. By means of the Hurst exponent, we investigate the kind of autocorrelation which is present and we detected the presence of multifractality. The volatility of the NYSE Index show a behavior analogous to a Wiener process. On the other hand, the probability density function was adjusted by a symmetric Lévy distribuition with alpha = 1.47. We present the variance autoregressive ARCH and GARCH models. More specifically, we focus on the Markovian GARCH(1,1) model. This model has three control parameters. We show that, for the NYSE Index, the use of the time autocorrelation to determinate the set of control parameters is not the best choice. Instead, results much more reasonable are obtained if the standardized sixth moment is used, as can be seen by the adjust of the time autocorrelation function. The proposal of the sixth moment is robust and applies for both the Gaussian and the Exponential GARCH models. We developed a series expansion technique to get the standardized sixth moment as a function of the three control parameters. We found an exact analytic expression for the kurtosis of the Exponential GARCH model. Both the Gaussian and the Exponential versions exhibit an equivalent performance in the description of the probability density function and the time autocorrelation function. However, with respect to the time scaling laws (measured by the probability of return to the origin) the Exponential model shows, in a clear and unequivocal way, a better performance than the Gaussian model, since it gives a time horizon exponent much more close to the real NYSE exponent.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第23541号 / 情博第771号 / 新制||情||132(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 河原 達也, 教授 黒橋 禎夫, 教授 森 信介 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

Former researches provide evidence that invasive species could alter ecosystem’s components, threaten native species and cause economic losses in southern forest lands. The objective of the project is to explore significant driving factors and develop geospatial models for monitoring, predicting and mapping the extent and conditions of major invasive species. In the study area, 16 invasive species were classified into four groups: regionally spreading species, regionally establishing species, locally spreading species and regionally colonizing species by population size and spatial characteristics. According to local Moran’s I, spatial autocorrelation existed in 16 invasive species. Autologistic model and simultaneous autoregressive model were employed to explore the relationships between spatial distribution and a set of indentified variables for Chinese privet, kudzu, Nepalese browntop and tallow tree at plot and county levels. The project showed that human-caused disturbances and forest types were significantly related to the spatial distribution of four invasive species in different scales.

This dissertation investigates the interactive or joint influence of autocorrelative processes (autoregressive-AR, moving average-MA, and autoregressive moving average-ARMA) and sample weights present in a longitudinal panel data set. Specifically, to what extent are the sample estimates influenced when autocorrelation (which is usually present in a panel data having correlated observations and errors) and sample weights (complex sample design feature used in longitudinal data having multi-stage sampling design) are modeled versus when they are not modeled or either one of them is taken into account. The current study utilized a Monte Carlo simulation design to vary the type and magnitude of autocorrelative processes and sample weights as factors incorporated in growth or latent curve models to evaluate the effect on sample latent curve estimates (mean intercept, mean slope, intercept variance, slope variance, and intercept slope correlation). Various latent curve models with weights or without weights were specified with an autocorrelative process and then fitted to data sets having either the AR, MA or ARMA process. The relevance and practical importance of the simulation results were ascertained by testing the joint influence of autocorrelation and weights on the Early Childhood Longitudinal Study for Kindergartens (ECLS-K) data set which is a panel data set having complex sample design features. The results indicate that autocorrelative processes and weights interact with each other as sources of error to a statistically significant degree. Accounting for just the autocorrelative process without weights or utilizing weights while ignoring the autocorrelative process may lead to bias in the sample estimates particularly in large-scale datasets in which these two sources of error are inherently embedded. The mean intercept and mean slope of latent curve models without weights was consistently underestimated when fitted to data sets having AR, MA or ARMA process. On the other hand, the intercept variance, intercept slope, and intercept slope correlation were overestimated for latent curve models with weights. However, these three estimates were not accurate as the standard errors associated with them were high. In addition, fit indices, AR and MA estimates, parsimony of the model, behavior of sample latent curve estimates, and interaction effects between autocorrelative processes and sample weights should be assessed for all the models before a particular model is deemed as most appropriate. If the AR estimate is high and MA estimate is low for a LCAR model than the other models that are fitted to a data set having sample weights and the fit indices are in the acceptable cut-off range, then the data set has a higher likelihood of having an AR process between the observations. If the MA estimate is high and AR estimate is low for a LCMA model than the other models that are fitted to a data set having sample weights and the fit indices are in the acceptable cut-off range, then the data set has a higher likelihood of having an MA process between the observations. If both AR and MA estimates are high for a LCARMA model than the other models that are fitted to a data set having sample weights and the fit indices are in the acceptable cut-off range, then the data set has a higher likelihood of having an ARMA process between the observations. The results from the current study recommends that biases from both autocorrelation and sample weights needs to be simultaneously modeled to obtain accurate estimates. The type of autocorrelation (AR, MA or ARMA), magnitude of autocorrelation, and sample weights influences the behavior of estimates and all the three facets should be carefully considered to correctly interpret the estimates especially in the context of measuring growth or change in the variable(s) of interest over time in large-scale longitudinal panel data sets.

No description available.

No description available.