一個社會經濟學的目標變量，經常存在兩種不同收集頻率的數據。由於較低頻率的一組數據通常由大型普查中所獲得，其準確度及可靠性會較高。因此較低頻率的一組數據一般會視作基準，用作對頻率較高的另一組數據進行修正。 / 在基準修正過程中，一般會假設調查誤差及目標數據的大小互相獨立，即「累加模型」。然而，現實中兩者通常是相關的，目標變量越大，調查誤差亦會越大，即「乘積模型」。對此問題，陳兆國及胡家浩提出了利用準線性回歸手法對乘積模型進行基準修正。在本論文中，假設調查誤差服從AR(1)模型，首先我們會示範如何利用準線性回歸手法及默認調查誤差模型進行基準數據修正。然後，運用基準預測的方式，提出一個對調查誤差模型的估計辦法。最後我們會比較兩者的表現以及一些選擇誤差模型的指引。 / For a target socio-economic variable, two sources of data with different collecting frequencies may be available in survey data analysis. In general, due to the difference of sample size or the data source, two sets of data do not agree with each other. Usually, the more frequent observations are less reliable, and the less frequent observations are much more accurate. In benchmarking problem, the less frequent observations can be treated as benchmarks, and will be used to adjust the higher frequent data. / In the common benchmarking setting, the survey error and the target variable are always assumed to be independent (Additive case). However, in reality, they should be correlated (Multiplicative case). The larger the variable, the larger the survey error. To deal with this problem, Chen and Wu (2006) proposed a regression method called quasi-linear regression for the multiplicative case. In this paper, by assuming the survey error to be an AR(1) model, we will demonstrate the benchmarking procedure using default error model for the quasi-linear regression. Also an error modelling procedure using benchmark forecast method will be proposed. Finally, we will compare the performance of the default error model with the fitted error model. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Luk, Wing Pan. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 56-57). / Abstracts also in Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Recent Development For Benchmarking Methods --- p.2 / Chapter 1.2 --- Multiplicative Case And Benchmarking Problem --- p.3 / Chapter 2 --- Benchmarking With Quasi-linear Regression --- p.8 / Chapter 2.1 --- Iterative Procedure For Quasi-linear Regression --- p.9 / Chapter 2.2 --- Prediction Using Default Value φ --- p.16 / Chapter 2.3 --- Performance Of Using Default Error Model --- p.17 / Chapter 3 --- Estimation Of φ Via BM Forecasting method --- p.26 / Chapter 3.1 --- Benchmark Forecasting Method --- p.26 / Chapter 3.2 --- Performance Of Benchmark Forecasting Method --- p.28 / Chapter 4 --- Benchmarking By The Estimated Value --- p.34 / Chapter 4.1 --- Benchmarking With The Estimated Error Model --- p.35 / Chapter 4.2 --- Performance Of Using Estimated Error Model --- p.36 / Chapter 4.3 --- Suggestions For Selecting Error Model --- p.45 / Chapter 5 --- Fitting AR(1) Model For Non-AR(1) Error --- p.47 / Chapter 5.1 --- Settings For Non-AR(1) Model --- p.47 / Chapter 5.2 --- Simulation Studies --- p.48 / Chapter 6 --- An Illustrative Example: The Canada Total Retail Trade Se-ries --- p.50 / Chapter 7 --- Conclusion --- p.54 / Bibliography --- p.56

Hung, King Chi. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2008. / Includes bibliographical references (leaves 78-52). / Abstracts in English and Chinese. / Abstract --- p.i / Acknowledgement --- p.iii / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Non-linear Time Series Models and Their Applications --- p.2 / Chapter 1.2 --- The SETAR Model --- p.4 / Chapter 1.3 --- Objectives and Organization of the Thesis --- p.6 / Chapter 2 --- The SETAR Model and Robust Test for Non-linearity --- p.8 / Chapter 2.1 --- A Brief Review of Existing Tests for Threshold-type Non-linearity --- p.9 / Chapter 2.2 --- Robust Tests for Threshold-type Non-linearity --- p.11 / Chapter 2.2.1 --- Tsay´ةs F Test --- p.12 / Chapter 2.2.2 --- The Proposed Test --- p.15 / Chapter 2.3 --- The Choice of the ψ-function --- p.23 / Chapter 2.4 --- A Simulation Study --- p.26 / Chapter 2.4.1 --- Data Generation Process (DGP) --- p.26 / Chapter 2.4.2 --- Simulation Findings --- p.29 / Chapter 3 --- Robust Estimation and Asymptotic Properties --- p.34 / Chapter 3.1 --- Least Squares Estimation --- p.37 / Chapter 3.2 --- Robust Estimation --- p.38 / Chapter 3.2.1 --- Asymptotic Properties --- p.40 / Chapter 3.3 --- A Simulation Study --- p.52 / Chapter 3.3.1 --- Data Generation Process (DGP) --- p.53 / Chapter 3.3.2 --- Simulation Findings --- p.55 / Chapter 3.3.3 --- Objective Function over r --- p.56 / Chapter 4 --- Numerical Example --- p.67 / Chapter 4.1 --- Methodology --- p.68 / Chapter 4.2 --- ASEAN Background --- p.69 / Chapter 4.2.1 --- Non-linearity tests on ASEAN Exchange Rate --- p.72 / Chapter 4.2.2 --- Estimation of the Return of Singaporean Dollar --- p.73 / Chapter 5 --- Conclusions and Further Research --- p.76 / References --- p.78

For the prediction of chaotic time series, a dichotomy has arisen between local approaches and global approaches. Local approaches hold the reputation of simplicity and feasibility, but they generally do not produce a compact description of the underlying system and are computationally intensive. Global approaches have the advantage of requiring less computation and are able to yield a global representation of the studied time series. However, due to the complexity of the time series process, it is often not easy to construct a global model to perform the prediction precisely. In addition to these approaches, a combination of the global and local techniques, called mixture of experts (ME), is also possible, where a smaller number of models work cooperatively to implement the prediction. This thesis reports on research about ME models for chaotic time series prediction. Based on a review of the techniques in time series prediction, a HMM-based ME model called "Time-line" Hidden Markov Experts (THME) is developed, where the trajectory of the time series is divided into some regimes in the state space and regression models called local experts are applied to learn the mapping on the regimes separately. The dynamics for the expert combination is a HMM, however, the transition probabilities are designed to be time-varying and conditional on the "real time" information of the time series. For the learning of the "time-line" HMM, a modified Baum-Welch algorithm is developed and the convergence of the algorithm is proved. Different versions of the model, based on MLP, RBF and SVM experts, are constructed and applied to a number of chaotic time series on both one-step-ahead and multi-step-ahead predictions. Experiments show that in general THME achieves better generalization performance than the corresponding single models in one-step-ahead prediction and comparable to some published benchmarks in multi-step-ahead prediction. Various properties of THME, such as the feature selection for trajectory dividing, the clustering techniques for regime extraction, the "time-line" HMM for expert combination and the performance of the model when it has different number of experts, are investigated. A number of interesting future directions for this work are suggested, which include the feature selection for regime extraction, the model selection for transition probability modelling, the extension to distribution prediction and the application on other time series.

Thesis (M. Phil.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Thesis (Ph. D.)--Ohio State University, 1982. / Includes vita. Includes bibliographical references (leaves 125-133). Available online via OhioLINK's ETD Center.

This thesis develops a novel robust periodogram method for detecting long memory. Though many test for long memory are based on the idea of linear regression, there exists no results in statistical literature on utilizing the robust regression methodology for detection of long memory. The advantage of the robust regression is a substantially less sensitivity to atypical observations or outliers, compared to the classical regression that is based on the least squares method. The thesis suggests two versions of the robust periodogram methods based on the least quan- tile and the least trimmed methods. The new robust periodogram methods are shown to provide smaller bias in long memory estimation when compared with the classical periodogram method. However, variability of estimation is increased. Therefore, we develop the bootstrapped modification of the new robust periodogram methods to reduce variability of estimation. The new bootstrapped modi¯cations of the robust periodogram tests substantially reduce variance of estimation and provides a competitively low bias. All proposed robust methods are illustrated by simulations and the case studies on currency exchange rates, and comparative analysis with other existing tests for long memory is carried out.

This thesis develops a novel robust periodogram method for detecting long memory. Though many test for long memory are based on the idea of linear regression, there exists no results in statistical literature on utilizing the robust regression methodology for detection of long memory. The advantage of the robust regression is a substantially less sensitivity to atypical observations or outliers, compared to the classical regression that is based on the least squares method. The thesis suggests two versions of the robust periodogram methods based on the least quan- tile and the least trimmed methods. The new robust periodogram methods are shown to provide smaller bias in long memory estimation when compared with the classical periodogram method. However, variability of estimation is increased. Therefore, we develop the bootstrapped modification of the new robust periodogram methods to reduce variability of estimation. The new bootstrapped modi¯cations of the robust periodogram tests substantially reduce variance of estimation and provides a competitively low bias. All proposed robust methods are illustrated by simulations and the case studies on currency exchange rates, and comparative analysis with other existing tests for long memory is carried out.

The effect on urban rainfall of pollution aerosols is studied both by data analysis and computational simulation. Our study examines data for urban areas undergoing decadal clean-up. We compare the annual precipitation between polluted sites and relatively clean sites through the time range before and during their clean-up periods to see how the air quality may affect the precipitation amount. By comparing the annual precipitation amount between two polluted sites with different elevations we demonstrate the role that elevation may play in rainfall suppression. Based on the data we collected, we built a model to analyze the relationship between air pollution aerosols and precipitation. Finally, we used a model of time dependent condensational aerosol growth to numerically study the relationship of air pollution aerosols and precipitation amount. Based on these results, we found a negative relationship of precipitation amount and air pollution amount; also, the simulation results clearly demonstrated that too many air pollution particles will deplete the water vapor and suppress further growth of condensation nuclei (CN) toward cloud condensation nuclei (CNN). This study supported the theoretical explanation on why air pollution could suppress urban rainfall.

The purpose of this exploratory case study was to describe an expert teacher’s decision-making system during interactive instruction using teacher self-report information, classroom observation data, and physiological recordings. Timed recordings of instructional interaction variables using an adapted Stallings Observation System were combined with simultaneous skin voltage measurements in time series analyses to describe observable and physiological elements of an expert teacher’s decision-making process. The mean and standard deviation of observable decision-action rates on teacher-identified “teaching days” were higher than the rates on “guiding” days. Bivariate time series analysis of decision-action rates and physiological response rates showed a significant positive relationship between the teacher’s decision-action rate and her physiological response rate on one teaching day. The positive relationship between the teacher’s decision-action rate and her physiological response rate was found to be context-dependent and related to the teaching strategy being used. High decision-action rates during direct instruction were associated with high physiological response rates compared to lower decision-action rates and physiological response rates while monitoring independent seatwork during a test. Correlation analysis of physiological response rates with time revealed slight, but statistically significant negative trends for four of the five observation days. Major features of the teacher’s decision-making system included focusing attention on academic instruction with the use of routines for managing students and materials to perform teaching tasks; both proactive and reactive improvisational decisions; and physiological events characteristic of autonomic nervous system activity during instructional sequences of high teacher-student interactivity. Damasio’s Somatic Marker Hypothesis (Damasio, 1999) is offered as an explanation for the generation of specific characteristics of the expert teacher’s instruction, such as the high frequency of decision-actions and automaticity of appropriate decisions.

The effect on urban rainfall of pollution aerosols is studied both by data analysis and computational simulation. Our study examines data for urban areas undergoing decadal clean-up. We compare the annual precipitation between polluted sites and relatively clean sites through the time range before and during their clean-up periods to see how the air quality may affect the precipitation amount. By comparing the annual precipitation amount between two polluted sites with different elevations we demonstrate the role that elevation may play in rainfall suppression. Based on the data we collected, we built a model to analyze the relationship between air pollution aerosols and precipitation. Finally, we used a model of time dependent condensational aerosol growth to numerically study the relationship of air pollution aerosols and precipitation amount. Based on these results, we found a negative relationship of precipitation amount and air pollution amount; also, the simulation results clearly demonstrated that too many air pollution particles will deplete the water vapor and suppress further growth of condensation nuclei (CN) toward cloud condensation nuclei (CNN). This study supported the theoretical explanation on why air pollution could suppress urban rainfall.