近年，プローブカーシステムによる交通情報収集が各地で進められている．プローブカーデータから生成される交通情報には，車両間での走行特性の違いやGPSエラー等によりばらつきが含まれる．このため，信頼性の高い情報を生成するためには多くのサンプル数が必要となる．このため，各リンクで必要なプローブサンプル数を効率よく収集するためには，プローブカーを適切に配備しなければならない．本研究は，名古屋市周辺で収集されたタクシープローブカーデータを用い，各道路リンクでの通過頻度や必要サンプル数をモデル化し，情報収集確率を最大とするプローブカーの最適割当て計画を行う．分析の結果，信頼性の高い情報を効率的に得るためには，より多くのプローブカーの確保や都心部を囲む広いエリアへの割当てが必要であることを示した．

textabstractOver the past decades, demographic changes, advances in transportation and communication technology, and the growth of the services sector have had a significant impact on the spatial structure of regions. Monocentric cities are disappearing and developing into polycentric metropolitan areas, while at the same time, social economic processes are taking place at an ever larger geographical scale, beyond that of the city, in which historically separate metropolitan areas are becoming increasingly functionally connected to form polycentric urban regions. Such urban networks are characterised by the lack of an urban hierarchy, a significant degree of spatial integration between different cities and, complementary relationships between centres, in that cities and towns have different economic specialisations. The growing literature on changing urban systems coincides with the increasing popularity of the urban network concept in contemporary spatial planning and policy, in which urban networks are often seen as a panacea for regional economic development problems. Polycentricity and spatial integration have become catchphrases, where polycentric development policies have been introduced to support territorial cohesion as well as higher levels of territorial competitiveness. Despite the enthusiasm for the ideas of a polycentric and networked spatial organisation, the assessment of the network concept leaves much to be desired. To what extent are regions becoming more polycentric and spatially integrated? Are relationships between cities in polycentric, spatially integrated regions complementary rather than competitive? And are polycentric, spatially integrated regions more economically efficient than their monocentric, non-integrated counterparts? In this study, these questions will be addressed.

textabstractA Guide to Modern Econometrics is a new textbook published by John Wiley and Sons. It covers a wide range of topics in applied econometrics in a concise and intuitive way. Some distinctive features: Emphasis on empirical relevance and intuition, paying attention to the links between alternative approaches. Limited use of matrix algebra. Coverage of many modern topics from time-series, cross-section and panel data econometrics. Concisely and carefully written, so that the reader does not get lost in the details. Full length empirical illustrations are provided throughout, typically taken from the modern economics literature and using full-size data sets. Empirical illustrations taken from finance, labour economics, environmental economics, monetary economics, international economics and many more. Exercises added to all chapters, with a focus on intuition and interpretation of results. Several exercises involve the use of actual data. Data sets used for illustrations and exercises are available from the internet.

his paper examines two alternative specifications for estimating event count models in which the data generating T process results in a larger number of zero counts than would be expected under standard distributional assumptions. I compare King’s "hurdle" event count model and Greene’s "zero-inflated" Poisson model, using data on Congressional responses to Supreme Court decisions from 1979 to 1988. I show that each of these models is a special case of a more general dual regime data generating process which results in extra-Poisson zero counts. Furthermore, because this data generating process can produce overdispersion in its own right, these models are also shown to be related to "variance function" negative binomial specifications. The underlying correspondence between these models leads to similar results in estimating and interpreting them in practice.

espanolEsta tesis doctoral propuso varios enfoques estadisticos para analizar la dinamica urbana con el objetivo de proporcionar herramientas para los procesos de toma de decisiones y estudios urbanos. Supuso que la actividad humana y la movilidad humana componen dinamicas urbanas. Inicialmente, estudio datos de redes sociales geolocalizados y los considero como un proxy de donde y cuando las personas realizan lo que se define como la actividad humana. Se emplearon tecnicas asociadas con modelos lineales generalizados, analisis de datos funcionales, agrupacion jerarquica y datos epidemicos, para explicar la distribucion espacio-temporal de los lugares donde las personas interactuan con sus redes sociales. Posteriormente, para comprender la movilidad en entornos urbanos, se utilizaron datos provenientes de un sistema de ferrocarril subterraneo. La informacion se considero medidas diarias repetidas para capturar la regularidad del comportamiento humano. Mediante la implementacion de metodos de componentes principales funcionales, analisis de datos y agrupacion jerarquica, fue posible describir el sistema e identificar patrones de movilidad humana. EnglishThis doctoral dissertation proposed several statistical approaches to analyse urban dynamics with aiming to provide tools for decision making processes and urban studies. It assumed that human activity and human mobility compose urban dynamics. Initially, it studied geolocated social media data and considered them as a proxy for where and when people carry out what it is defined as the human activity. It employed techniques associated with generalised linear models, functional data analysis, hierarchical clustering, and epidemic data, to explain the spatio-temporal distribution of the places where people interact with their social networks. Afterwards, to understand the mobility in urban environments, data coming from an underground railway system were used. The information was considered repeated daily measurements to capture the regularity of human behaviour. By implementing methods from functional principal components data analysis and hierarchical clustering, it was possible to describe the system and identify human mobility patterns.

cientific knowledge is an important ingredient in the innovation process. Drawing on the knowledge-based view of the firm and the literature on the relationship between science and technology, this paper scrutinizes the importance of university scientists mobility for firms innovative activities. Combining patent data and matched employer-employee data for Danish firms, we can track the labor mobility of R&D workers from 1999 to 2004. We find that new joiners contribute more than long-term employees to innovative activity in the focal firm. Among new firm recruits, we observe that newly hired former university researchers contribute more to innovative activity than newly hired recent graduates or joiners from firms, but only in firms with a high level of absorptive capacity in the form of recent experience of hiring university researchers. We find also that firms recent experience of hiring university researchers enhances the effect of newly hired recent graduates contributions to innovation.

Zero-inflated count outcomes arise quite often in research and practice. Parametric models such as the zero-inflated Poisson and zero-inflated negative binomial are widely used to model such responses. Like most parametric models, they are quite sensitive to departures from assumed distributions. Recently, new approaches have been proposed to provide distribution-free, or semi-parametric, alternatives. These methods extend the generalized estimating equations to provide robust inference for population mixtures defined by zero-inflated count outcomes. In this paper, we propose methods to extend smoothly clipped absolute deviation (SCAD)-based variable selection methods to these new models. Variable selection has been gaining popularity in modern clinical research studies, as determining differential treatment effects of interventions for different subgroups has become the norm, rather the exception, in the era of patent-centered outcome research. Such moderation analysis in general creates many explanatory variables in regression analysis, and the advantages of SCAD-based methods over their traditional counterparts render them a great choice for addressing this important and timely issues in clinical research. We illustrate the proposed approach with both simulated and real study data. Copyright © 2016 John Wiley & Sons, Ltd.

Zero-inflated Poisson (ZIP) and negative binomial (ZINB) models are widely used to model zero-inflated count responses. These models extend the Poisson and negative binomial (NB) to address excessive zeros in the count response. By adding a degenerate distribution centered at 0 and interpreting it as describing a non-risk group in the population, the ZIP (ZINB) models a two-component population mixture. As in applications of Poisson and NB, the key difference between ZIP and ZINB is the allowance for overdispersion by the ZINB in its NB component in modeling the count response for the at-risk group. Overdispersion arising in practice too often does not follow the NB, and applications of ZINB to such data yield invalid inference. If sources of overdispersion are known, other parametric models may be used to directly model the overdispersion. Such models too are subject to assumed distributions. Further, this approach may not be applicable if information about the sources of overdispersion is unavailable. In this paper, we propose a distribution-free alternative and compare its performance with these popular parametric models as well as a moment-based approach proposed by Yu et al. [Statistics in Medicine 2013; 32: 2390-2405]. Like the generalized estimating equations, the proposed approach requires no elaborate distribution assumptions. Compared with the approach of Yu et al., it is more robust to overdispersed zero-inflated responses. We illustrate our approach with both simulated and real study data.

Within general insurance, pricing of premiums is always a challenging task. Frequency of claims plays a big part in pricing of premiums. Frequency of claims are determined by the attributes of a particular policy holder. Count regression analysis allows one to nd out which characteristic of a policy holder plays a signi cant role in determining the frequency of claim and also in predicting the frequency of claims given the characteristics of a particular policy holder. The objective of this thesis is to nd out which among the Poisson, NB1 and NB2 models is a better t to the count data under consideration. The count data is from Kenendia insurance. The best model is chosen based on the log-likelihood method and the Akaike’s Information Criteria (AIC).

With the growth of online shopping coupled with mobile technology, user-generated product reviews have become an important source of information for product diagnosticity. A significant academic endeavor has been made to comprehend what information factors of reviews help prospective customers better diagnose products. One such factor is review depth that is estimated by the number of a review’s words. We propose review breadth as an additional factor based on a review’s number of topics—the more review breadth, the more diverse information. By conducting the statistical and predictive analyses, we demonstrate that review breadth reliably measures a review’s information. This study makes academic and practical contributions. For academic researchers, review breadth is worth considering as a factor to estimate a review’s information over and above review depth. Based on the two information factors of review breadth and review depth, practitioners can recommend more helpful product reviews to their prospective customers.

While the course of technological change is widely accepted to be highly uncertain and unpredictable, little work has identified or studied the ultimate sources and causes of that uncertainty. This paper proposes that purely technological uncertainty derives from inventors’ search processes with unfamiliar components and component combinations. Experimentation with new components and new combinations leads to less useful inventions on average, but it also implies an increase in the variability that can result in both failure and breakthrough. Negative binomial count and dispersion models with patent citation data demonstrate that new combinations are indeed more variable. In contrast to predictions, however, the reuse of components has a nonmonotonic and eventually positive effect on variability. (Invention; Search; Recombination; Negative Binomial Dispersion Model)

While the course of technological change is widely accepted to be highly uncertain and unpredictable, little work has identified or studied the ultimate sources and causes of that uncertainty. This paper proposes that purely technological uncertainty derives from inventors' search processes with unfamiliar components and component combinations. Experimentation with new components and new combinations leads to less useful inventions on average, but it also implies an increase in the variability that can result in both failure and breakthrough. Negative binomial count and dispersion models with patent citation data demonstrate that new combinations are indeed more variable. In contrast to predictions, however, the reuse of components has a nonmonotonic and eventually positive effect on variability.

While rail has been the focus of most planning for Transit Oriented Development (TOD), there has been much interest in bus-related TOD with an emphasis on new bus rapid transit (BRT) systems in North and South America and Australia. This paper takes a critical look at the strengths and challenges of bus-based transit systems compared to rail in relation to TOD. It includes a literature review and an assessment of TOD-related developments. The focus is on performance of BRT systems in relation to TOD with specific reference to BRT systems in Australia. In addition, TOD related to local suburban bus service is examined. The paper describes the general concept of TOD and how it relates to features of transit modes, outlines the literature relevant to bus-based TOD, and identifies the strengths and challenges of bus-based transit systems in relation to TOD. It concludes by summarizing the relative strengths and challenges of BRT and local bus services compared to rail. The findings of the review are used to identify ways in which bus-based TOD might be better planned and implemented.

When scientists move to a new location, they may acquire and recombine new knowledge and form new collaborative networks, which can benefit their innovative productivity. However, the extent of this benefit may be conditional upon how entrenched these scientists were in their former locations and networks. In this study, we investigate the impact of scientific mobility and entrenchment on innovative productivity, in terms of the number of patents applied. Drawing upon the population of 5,809 scientists who have contributed to innovation in genomics by being awarded at least one gene patent, we identify focal scientists who have moved as a result of an exogenous shock from the 1994 Northridge earthquake vis-a-vis similar scientists who did not move. Our difference-in-differences estimates show that these focal scientists who have moved experience higher innovative productivity post move relative to non-movers. However, we also find that this effect is negatively influenced by the presence of a majority of collaborators who were collocated with the focal scientist in the pre-move location and the length of the scientist's tenure, while it is positively influenced by the presence of a recent relocation by the focal scientist into the pre-move location. These findings highlight the previously underexplored disruptive effects of mobility on knowledge acquisition, assimilation and innovative productivity, and have important policy and strategy implications.

When econometric models are estimated by maximum likelihood, the conditional information matrix variance estimator is usually avoided in choosing a method for estimating the variance of the parameter estimate. However, the conditional information matrix estimator attains the semiparametric efficiency bound for the variance estimation problem. Unfortunately, for even moderately complex models, the integral involved in computation of the conditional information matrix estimator is prohibitively difficult to solve. Simulation is suggested to approximate the integral, and two simulation variance estimators are proposed. Monte Carlo results suggest these estimators are attractive in providing accurate confidence interval coverage rates compared to the standard maximum likelihood variance estimators.

When agents originate and develop inventions, new discoveries, or scientific breakthroughs, do all residual cash flows go to principals - the risk bearers who own the infrastructures, systems, and the productive assets? This research addresses this question, focusing on university scientists who essentially bypass their institutions when they privately sell or license discoveries made at laboratories of said institutions directly to the market. Using a random sample of 54 U.S. universities and 23 394 faculty members, the study shows that bypassing activity is reduced when universities rely on autonomous technology licensing offices (TLOs) and when faculty departments receive greater shares of the royalties from the licensing of said discoveries. Conversely, bypassing activity is increased with more valuable discoveries and heightened entrepreneurial activities on university campuses.

What influences the Northern media's coverage of events and abuses in explicit human rights terms? Do international NGOs have an impact, and, if so, when are they most effective? This article addresses these questions with regression analysis of human rights reporting by The Economist and Newsweek from 1986 to 2000, covering 145 countries. First, it finds that these two media sources cover abuses in human rights terms more frequently when they occur in countries with higher levels of state repression, economic development, population, and Amnesty International attention. There is also some evidence that political openness, number of battle-deaths, and civil societies affect coverage, although these effects were not robust. Second, it finds that Amnesty International's press releases appear to have less impact on media coverage when discussing abuses in countries that are central to the media's zone of concern. Indeed, Amnesty's press advocacy may be more effective when addressing violations in lesser-noticed countries. The article attributes this to the saturation of coverage of abuses in highly mediatized countries. Cumulative attention by multiple journalists and others raises a country's media profile but also makes it more difficult for any one voice to be heard. The authors conclude that Amnesty's press advocacy may have greater media impact when focusing on abuses in countries located away from the media's core areas of concern. Overall, the authors are encouraged by the Northern media's sensitivity to actual patterns of repression and to Amnesty's lobbying, since both indicate that the media is potentially a useful ally in efforts to combat abuses worldwide. Yet, the discouraging effects of poverty on the media's human rights coverage are cause for concern.

We tested the block-level deterrent effects on crime of uniformed police raids of crack houses. Court-authorized raids were legally possible on 207 blocks with at least five calls for police service in the preceding 30 days. Raids were assigned randomly to 104 locations and were conducted at 98 of those sites; the other 109 were left alone. Experimental blocks, in relation to controls, showed reductions in both calls for service and offense reports, but effects were quite small and decayed in two weeks. Raids in which arrests were made (23 of 104 assigned) had no consistently different impact from raids in which no arrests were made. Raids had more effect on calls for service in the winter than in the spring, but we found little seasonal or period difference in effects of raids on offense reports. Alternative police methods may be far more cost-effective than raids in “harm reduction” for crack houses.

We study the link between family violence and the emotional cues associated with wins and losses by professional football teams. We hypothesize that the risk of violence is affected by the “gain-loss” utility of game outcomes around a rationally expected reference point. Our empirical analysis uses police reports of violent incidents on Sundays during the professional football season. Controlling for the pregame point spread and the size of the local viewing audience, we find that upset losses (defeats when the home team was predicted to win by four or more points) lead to a 10% increase in the rate of at-home violence by men against their wives and girlfriends. In contrast, losses when the game was expected to be close have small and insignificant effects. Upset wins (victories when the home team was predicted to lose) also have little impact on violence, consistent with asymmetry in the gain-loss utility function. The rise in violence after an upset loss is concentrated in a narrow time window near the end of the game and is larger for more important games. We find no evidence for reference point updating based on the halftime score.

We study the effect of the tie strength of inter-firm R&D partnerships on the innovation performance of companies in four high-tech sectors (pharmaceuticals, computers, semi-conductors and telecom). Returning to Granovetter's (1973) seminal contribution, tie strength is analysed through a broad multi-dimensional perspective. We find that inter-firm R&D network ties that are stronger in terms of their extent (measured by the length and multitude of R&D partnerships) and weaker in terms of their depth (the degree of cooperation and the similarity of ties of companies) improve the innovation performance of companies. Interestingly, we find strong support for the role of these R&D ties in the context of the run-of-the-mill innovation performance of companies. However, there is no support for this effect on significant innovations of companies.