In this paper, we briefly describe the automatic model selection which is provided by Autometrics in the PcGive program. The modeler only needs to specify the initial model and the significance level at which to reduce the model. Then, the algorithm does the rest. The properties of Autometrics are discussed. We also explain its background concepts and try to see whether the model selected by the Autometrics can perform well. For a given data set, we use Autometrics to find a “new” model, and then compare the “new” model with a previously selected one by another modeler. It is an interesting issue to see whether Autometrics can also find models which fit better to the given data. As an illustration, we choose three examples. It is true that Autometrics is labor saving and always gives us a parsimonious model. It is really an invaluable instrument for social science. But, we still need more examples to strongly support the idea that Autometrics can find a model which fits the data better, just a few examples in this paper is far from enough.

In many high dimensional problems, the dependence structure among the variables can be quite complex. An appropriate use of the regularization techniques coupled with other classical statistical methods can often improve estimation and prediction accuracy and facilitate model interpretation, by seeking a parsimonious model representation that involves only the subset of revelent variables. We propose two regularized stochastic regression approaches, for efficiently estimating certain sparse dependence structure in the data. We first consider a multivariate regression setting, in which the large number of responses and predictors may be associated through only a few channels/pathways and each of these associations may only involve a few responses and predictors. We propose a regularized reduced-rank regression approach, in which the model estimation and rank determination are conducted simultaneously and the resulting regularized estimator of the coefficient matrix admits a sparse singular value decomposition (SVD). Secondly, we consider model selection of subset autoregressive moving-average (ARMA) modelling, for which automatic selection methods do not directly apply because the innovation process is latent. We propose to identify the optimal subset ARMA model by fitting a penalized regression, e.g. adaptive Lasso, of the time series on its lags and the lags of the residuals from a long autoregression fitted to the time-series data, where the residuals serve as proxies for the innovations. Computation algorithms and regularization parameter selection methods for both proposed approaches are developed, and their properties are explored both theoretically and by simulation. Under mild regularity conditions, the proposed methods are shown to be selection consistent, asymptotically normal and enjoy the oracle properties. We apply the proposed approaches to several applications across disciplines including cancer genetics, ecology and macroeconomics.

This dissertation contains three independent essays relating energy economics. The first essay investigates price asymmetry of diesel in South Korea by using the error correction model. Analyzing weekly market prices in the pass-through of crude oil, this model shows asymmetric price response does not exist at the upstream market but at the downstream market. Since time-variant residuals are found by the specified models for both weekly and daily retail prices at the downstream level, these models are implemented by a Generalized Autoregressive Conditional Heteroskedasticity (GARCH) process. The estimated results reveal that retail prices increase fast in the rise of crude oil prices but decrease slowly in the fall of those. Surprisingly, retail prices rarely respond to changes of crude oil prices for the first five days. Based on collusive behaviors of retailers, this price asymmetry in Korea diesel market is explained. The second essay aims to evaluate the new incentive system for biodiesel in South Korea, which keeps the blend mandate but abolishes tax credits for government revenues. To estimate changed welfare from the new policy, a multivariate stochastic simulation method is applied into time-series data for the last five years. From the simulation results, the new biodiesel policy will lead government revenues to increases with the abolishment of tax credit. However, increased prices of blended diesel will cause to decrease demands of both biodiesel and blended diesel, so consumer and producer surplus in the transport fuel market will decrease. In the third essay, the Regression - Seasonal Autoregressive Integrated Moving Average (REGSARIMA) model is employed to predict the impact of air temperature on daily peak load demand in Houston. Compared with ARIMA and Seasonal Model, a REGARIMA model provides the more accurate prediction for daily peak load demand for the short term. The estimated results reveal air temperature in the Houston areas causes an increase in electricity consumption for cooling but to save that for heating. Since the daily peak electricity consumption is significantly affected by hot air temperature, this study makes a conclusion that it is necessary to establish policies to reduce urban heat island phenomena in Houston.

This study aims to investigate short-to-medium forecasting and simulation of streamflow, water levels, and sediments in Canada using Seasonal Autoregressive Integrated Moving Average (SARIMA) time series models. The methodology can account for linear trends in the time series that may result from climate and environmental changes. A Universal Canadian forecast Application using python web interface was developed to generate short-term forecasts using SARIMA. The Akaike information criteria was used as performance criteria for generating efficient SARIMA models. The developed models were validated by analyzing the residuals. Several stations from the Canadian Hydrometric Database (HYDAT) displaying a linear upward or downward trend were identified to validate the methodology. Trends were detected using the Man-Kendall test. The Nash-Sutcliffe efficiency coefficients (Nash ad Sutcliffe, 1970) of the developed models indicate that they are acceptable. The models can be used for short term (1 to 7 days) and medium-term (7 days to six months) forecasting of streamflow, water levels and sediments at all Canadian hydrometric stations. Such a forecast can be used for water resources management and help mitigate the effects of floods and droughts. The models can also be used to generate long time-series that can be used to test the performance of water resources systems. Finally, we have automated the process of analysis, model-building and forecasting streamflow, water levels, and sediments by building a python-based application easily extendable and user-friendly. Therefore, automating the SARIMA calibration and forecasting process for all Canadian stations for the HYDAT database will prove to be a very useful tool for decision-makers and other entities in the field of hydrological study.

本論文採用了由Box and Jenkins(1976)所提出的ARIMA模型，以及由BOX and Tiao(1975)所提出的Intervention Model，去配適台灣的失業率型態，以及比較其預測的結果。 結果顯示出台灣的失業率具有季節性的型態，亦即台灣的失業率並非僅僅受到月分之間的相關，年分之間也有所關連。是故，當本論文在預測失業率的水準時，也考慮到此一因素，加入季節性的ARIMA模型對台灣的失業率加以預測。另外，時間序列的資料常常受到外生因素的干擾。對於失業率來說，政策上的改變將會影響失業率本身的結構，因此利用介入模式預測失業率，可以得到一組較精確的預測值。介入模式的事件有以下五個，分別是解嚴、六年國建、台灣引進外勞、中共飛彈試射、新十大建設。前四個事件的確影響了失業率的結構，不過第五項，也就是新十大建設並沒有顯著影響失業率的結構。理由可能是新十大建設的內容並不能合宜的解決經濟上與社會上的問題，以及這些建設尚未完工，以致無法達到期預期的效果。 比較兩模型的預測結果時，採用了MPE、MSE、MAE、MAPE作為模型評估的準則，結果指出介入模式的預測結果比起季節性ARIMA的預測結果來的有效率。 / This article adopts the ARIMA model, which was first introduced by Box and Jenkins (1976), and the intervention model, which was developed by Box and Tiao (1975), to fit the time series data for the unemployment rate in Taiwan, and thus to compare the results of the forecasts. The results reveal that there is a seasonal effect in the data on the unemployment rate. This indicates that the unemployment rate figures are not only related from month to month but are also related from year to year. When forecasting the level of unemployment, we should examine not only the neighboring months but also the corresponding months in the previous year. Time series are frequently affected by certain external events. In the discussion on the unemployment rate, the policies implemented by the government as well as military threats indeed influence the structure of the series. By making a forecast using the intervention model, we can evaluate the effect of the external events which would give rise to more accurate forecasts. In this study, there were five interventions included in relation to the unemployment rate series, which were as follows. First, the lifting of Martial Law in February 1987. Second, the Six-year National Development Plan launched in June 1991. Third, the hiring of foreign labor in Taiwan, which took effect in October 1991. Fourth, the threats of missile tests from the PRC in Feb 1996. Fifth, the ten new construction programs launched in November 2003. The first four events were indeed found to give rise to a structural change in the unemployment rate series at the moment when they occurred. This result might also have implied that not all of the actual effect of expansionary policies could have exactly decreased the unemployment rate, and therefore have solved the economic and social problems simultaneously. When we refer to the comparison of the above two models, the ultimate choice of a model may depend on its goodness of fit, such as the residual mean square, AIC, or BIC. As the main purpose of this study is to forecast future values, the alternative criteria for model selection can be based on forecast errors. The comparison is based on statistics such as MPE, MSE, MAE and MAPE. The results indicate that the intervention model outperforms the seasonal ARIMA model.