Showing papers on "Matching (statistics) published in 2006"

Large Sample Properties of Matching Estimators for Average Treatment Effects

Abstract: Matching estimators for average treatment effects are widely used in evaluation research despite the fact that their large sample properties have not been established in many cases. The absence of formal results in this area may be partly due to the fact that standard asymptotic expansions do not apply to matching estimators with a fixed number of matches because such estimators are highly nonsmooth functionals of the data. In this article we develop new methods for analyzing the large sample properties of matching estimators and establish a number of new results. We focus on matching with replacement with a fixed number of matches. First, we show that matching estimators are not N 1/2 -consistent in general and describe conditions under which matching estimators do attain N 1/2 -consistency. Second, we show that even in settings where matching estimators are N 1/2 -consistent, simple matching estimators with a fixed number of matches do not attain the semiparametric efficiency bound. Third, we provide a consistent estimator for the large sample variance that does not require consistent nonparametric estimation of unknown functions. Software for implementing these methods is available in Matlab, Stata, and R.

Correcting Sample Selection Bias by Unlabeled Data

Abstract: We consider the scenario where training and test data are drawn from different distributions, commonly referred to as sample selection bias. Most algorithms for this setting try to first recover sampling distributions and then make appropriate corrections based on the distribution estimate. We present a nonparametric method which directly produces resampling weights without distribution estimation. Our method works by matching distributions between training and testing sets in feature space. Experimental results demonstrate that our method works well in practice.

LSB matching revisited

Abstract: This letter proposes a modification to the least-significant-bit (LSB) matching, a steganographic method for embedding message bits into a still image. In the LSB matching, the choice of whether to add or subtract one from the cover image pixel is random. The new method uses the choice to set a binary function of two cover pixels to the desired value. The embedding is performed using a pair of pixels as a unit, where the LSB of the first pixel carries one bit of information, and a function of the two pixel values carries another bit of information. Therefore, the modified method allows embedding the same payload as LSB matching but with fewer changes to the cover image. The experimental results of the proposed method show better performance than traditional LSB matching in terms of distortion and resistance against existing steganalysis.

Teacher-Student Matching and the Assessment of Teacher Effectiveness

Abstract: Administrative data on fifth grade students in North Carolina shows that more highly qualified teachers tend to be matched with more advantaged students, both across schools and in many cases within them. This matching biases estimates of the relationship between teacher characteristics and achievement; we isolate this bias in part by focusing on schools where students are distributed relatively evenly across classrooms. Teacher experience is consistently associated with achievement; teacher licensure test scores associate with math achievement. These returns display a form of heterogeneity across students that may help explain why the observed form of teacher-student matching persists in equilibrium.

Automated semantic web service discovery with OWLS-MX

Abstract: We present an approach to hybrid semantic Web service matching that complements logic based reasoning with approximate matching based on syntactic IR based similarity computations. The hybrid matchmaker, called OWLS-MX, applies this approach to services and requests specified in OWL-S. Experimental results of measuring performance and scalability of different variants of OWLS-MX show that under certain constraints logic based only approaches to OWL-S service I/O matching can be significantly outperformed by hybrid ones.

Who marries whom and why

Abstract: This paper proposes and estimates a static transferable utility model of the marriage market. The model generates a nonparametric marriage matching function with spillover effects. It rationalizes the standard interpretation of marriage rate regressions and points out its limitations. The model was used to estimate U.S. marital behavior in 1971/72 and 1981/82. The marriage matching function estimates show that the gains to marriage for young adults fell substantially over the decade. Unlike contradictory marriage rate regression results, the marriage matching function estimates showed that the legalization of abortion had a significant quantitative impact on the fall in the gains to marriage for young men and women.

From Temporary Help Jobs to Permanent Employment: What Can We Learn from Matching Estimators and their Sensitivity?

Abstract: The diffusion of Temporary Work Agency (TWA) jobs originated a harsh policy debate and ambiguous empirical evidence. Results for the US, based on quasi-experimental evidence, suggest that a TWA assignment decreases the probability of finding a stable job, while results for Europe, based on the Conditional Independence Assumption (CIA), typically reach opposite conclusions. Using data for two Italian regions, we use a matching estimator to show that TWA assignments can be an effective springboard to permanent employment. We also propose a simulation-based sensitivity analysis, which highlights that only for one of these two regions our results are robust to specific failures of the CIA. We conclude that European studies based on the CIA should not be automatically discarded, but should be put under the scrutiny of a sensitivity analysis like the one we propose.

Optimal Full Matching and Related Designs via Network Flows

Abstract: In the matched analysis of an observational study, confounding on covariates X is addressed by comparing members of a distinguished group (Z = 1) to controls (Z = 0) only when they belong to the same matched set. The better matchings, therefore, are those whose matched sets exhibit both dispersion in Z and uniformity in X. For dispersion in Z, pair matching is best, creating matched sets that are equally balanced between the groups; but actual data place limits, often severe limits, on matched pairs' uniformity in X. At the other extreme is full matching, the matched sets of which are as uniform in X as can be, while often so poorly dispersed in Z as to sacrifice efficiency.This article presents an algorithm for exploring the intermediate territory. Given requirements on matched sets' uniformity in X and dispersion in Z, the algorithm first decides the requirements' feasibility. In feasible cases, it furnishes a match that is optimal for X-uniformity among matches with Z-dispersion as stipulated. To illus...

A comparison of propensity score methods: a case-study estimating the effectiveness of post-AMI statin use.

Abstract: There is an increasing interest in the use of propensity score methods to estimate causal effects in observational studies. However, recent systematic reviews have demonstrated that propensity score methods are inconsistently used and frequently poorly applied in the medical literature. In this study, we compared the following propensity score methods for estimating the reduction in all-cause mortality due to statin therapy for patients hospitalized with acute myocardial infarction: propensity-score matching, stratification using the propensity score, covariate adjustment using the propensity score, and weighting using the propensity score. We used propensity score methods to estimate both adjusted treated effects and the absolute and relative risk reduction in all-cause mortality. We also examined the use of statistical hypothesis testing, standardized differences, box plots, non-parametric density estimates, and quantile-quantile plots to assess residual confounding that remained after stratification or matching on the propensity score. Estimates of the absolute reduction in 3-year mortality ranged from 2.1 to 4.5 per cent, while estimates of the relative risk reduction ranged from 13.3 to 17.0 per cent. Adjusted estimates of the reduction in the odds of 3-year death varied from 15 to 24 per cent across the different propensity score methods.

How Smart is Smart Money? A Two-Sided Matching Model of Venture Capital

Abstract: I find that companies funded by more experienced VCs are more likely to go public. This follows both from the direct influence of more experienced VCs and from sorting in the market, which leads experienced VCs to invest in better companies. Sorting creates an endogeneity problem, but a structural model based on a Two-Sided Matching model is able to exploit the characteristics of the other agents in the market to separately identify and estimate influence and sorting. Both effects are found to be significant, but sorting is almost twice as important as influence for the difference in IPO rates.

Matching 2.5D face scans to 3D models

Abstract: The performance of face recognition systems that use two-dimensional images depends on factors such as lighting and subject's pose. We are developing a face recognition system that utilizes three-dimensional shape information to make the system more robust to arbitrary pose and lighting. For each subject, a 3D face model is constructed by integrating several 2.5D face scans which are captured from different views. 2.5D is a simplified 3D (x,y,z) surface representation that contains at most one depth value (z direction) for every point in the (x, y) plane. Two different modalities provided by the facial scan, namely, shape and texture, are utilized and integrated for face matching. The recognition engine consists of two components, surface matching and appearance-based matching. The surface matching component is based on a modified iterative closest point (ICP) algorithm. The candidate list from the gallery used for appearance matching is dynamically generated based on the output of the surface matching component, which reduces the complexity of the appearance-based matching stage. Three-dimensional models in the gallery are used to synthesize new appearance samples with pose and illumination variations and the synthesized face images are used in discriminant subspace analysis. The weighted sum rule is applied to combine the scores given by the two matching components. Experimental results are given for matching a database of 200 3D face models with 598 2.5D independent test scans acquired under different pose and some lighting and expression changes. These results show the feasibility of the proposed matching scheme.

A Comparison of Personal Name Matching: Techniques and Practical Issues

Abstract: Finding and matching personal names is at the core of an increasing number of applications: from text and Web mining, search engines, to information extraction, dedupli- cation and data linkage systems Variations and errors in names make exact string matching problematic, and ap- proximate matching techniques have to be applied When compared to general text, however, personal names have different characteristics that need to be considered In this paper we discuss the characteristics of personal names and present potential sources of variations and errors We then overview a comprehensive number of commonly used, as well as some recently developed name matching techniques Experimental comparisons using four large name data sets indicate that there is no clear best matching technique

The Beveridge Curve

Abstract: The Beveridge curve depicts a negative relationship between unemployed workers and job vacancies, a robust finding across countries. The position of the economy on the curve gives an idea as to the state of the labour market. The modern underlying theory is the search and matching model, with workers and firms engaging in costly search leading to random matching. The Beveridge curve depicts the steady state of the model, whereby inflows into unemployment are equal to the outflows from it, generated by matching.

Matching with Shape Contexts.

Abstract: We introduce a new shape descriptor, the shape context, for measuring shape similarity and recovering point correspondences. The shape context describes the coarse arrangement of the shape with respect to a point inside or on the boundary of the shape. We use the shape context as a vector-valued attribute in a bipartite graph matching framework. Our proposed method makes use of a relatively small number of sample points selected from the set of detected edges; no special landmarks or keypoints are necessary. Tolerance and/or invariance to common image transformations are available within our framework. Using examples involving both silhouettes and edge images, we demonstrate how the solution to the graph matching problem provides us with correspondences and a dissimilarity score that can be used for object recognition and similarity-based retrieval.

Statistical matching: theory and practice

Abstract: Preface. 1 The Statistical Matching Problem. 1.1 Introduction. 1.2 The Statistical Framework. 1.3 The Missing Data Mechanism in the Statistical Matching Problem. 1.4 Accuracy of a Statistical Matching Procedure. 1.4.1 Model assumptions. 1.4.2 Accuracy of the estimator. 1.4.3 Representativeness of the synthetic file. 1.4.4 Accuracy of estimators applied on the synthetic data set. 1.5 Outline of the Book. 2 The Conditional Independence Assumption. 2.1 The Macro Approach in a Parametric Setting. 2.1.1 Univariate normal distributions case. 2.1.2 The multinormal case. 2.1.3 The multinomial case. 2.2 The Micro (Predictive) Approach in the Parametric Framework. 2.2.1 Conditional mean matching. 2.2.2 Draws based on conditional predictive distributions. 2.2.3 Representativeness of the predicted files. 2.3 Nonparametric Macro Methods. 2.4 The Nonparametric Micro Approach. 2.4.1 Random hot deck. 2.4.2 Rank hot deck. 2.4.3 Distance hot deck. 2.4.4 The matching noise. 2.5 Mixed Methods. 2.5.1 Continuous variables. 2.5.2 Categorical variables. 2.6 Comparison of Some Statistical Matching Procedures under the CIA. 2.7 The Bayesian Approach. 2.8 Other IdentifiableModels. 2.8.1 The pairwise independence assumption. 2.8.2 Finite mixture models. 3 Auxiliary Information. 3.1 Different Kinds of Auxiliary Information. 3.2 Parametric Macro Methods. 3.2.1 The use of a complete third file. 3.2.2 The use of an incomplete third file. 3.2.3 The use of information on inestimable parameters. 3.2.4 The multinormal case. 3.2.5 Comparison of different regression parameter estimators through simulation. 3.2.6 The multinomial case. 3.3 Parametric Predictive Approaches. 3.4 Nonparametric Macro Methods. 3.5 The Nonparametric Micro Approach with Auxiliary Information. 3.6 Mixed Methods. 3.6.1 Continuous variables. 3.6.2 Comparison between some mixed methods. 3.6.3 Categorical variables. 3.7 Categorical Constrained Techniques. 3.7.1 Auxiliary micro information and categorical constraints. 3.7.2 Auxiliary information in the form of categorical constraints. 3.8 The Bayesian Approach. 4 Uncertainty in Statistical Matching. 4.1 Introduction. 4.2 A Formal Definition of Uncertainty. 4.3 Measures of Uncertainty. 4.3.1 Uncertainty in the normal case. 4.3.2 Uncertainty in the multinomial case. 4.4 Estimation of Uncertainty. 4.4.1 Maximum likelihood estimation of uncertainty in the multinormal case. 4.4.2 Maximum likelihood estimation of uncertainty in the multinomial case. 4.5 Reduction of Uncertainty: Use of Parameter Constraints. 4.5.1 The multinomial case. 4.6 Further Aspects of Maximum Likelihood Estimation of Uncertainty. 4.7 An Example with Real Data. 4.8 Other Approaches to the Assessment of Uncertainty. 4.8.1 The consistent approach. 4.8.2 The multiple imputation approach. 4.8.3 The de Finetti coherence approach. 5 Statistical Matching and Finite Populations. 5.1 Matching Two Archives. 5.1.1 Definition of the CIA. 5.2 Statistical Matching and Sampling from a Finite Population. 5.3 Parametric Methods under the CIA. 5.3.1 The macro approach when the CIA holds. 5.3.2 The predictive approach. 5.4 Parametric Methods when Auxiliary Information is Available. 5.4.1 The macro approach. 5.4.2 The predictive approach. 5.5 File Concatenation. 5.6 Nonparametric Methods. 6 Issues in Preparing for Statistical Matching. 6.1 Reconciliation of Concepts and Definitions of Two Sources. 6.1.1 Reconciliation of biased sources. 6.1.2 Reconciliation of inconsistent definitions. 6.2 How to Choose the Matching Variables. 7 Applications. 7.1 Introduction. 7.2 Case Study: The Social Accounting Matrix. 7.2.1 Harmonization step. 7.2.2 Modelling the social accounting matrix. 7.2.3 Choosing the matching variables. 7.2.4 The SAM under the CIA. 7.2.5 The SAM and auxiliary information. 7.2.6 Assessment of uncertainty for the SAM. A Statistical Methods for Partially Observed Data. A.1 Maximum Likelihood Estimation with Missing Data. A.1.1 Missing data mechanisms. A.1.2 Maximum likelihood and ignorable nonresponse. A.2 Bayesian Inference withMissing Data. B Loglinear Models. B.1 Maximum Likelihood Estimation of the Parameters. C Distance Functions. D Finite Population Sampling. E R Code. E.1 The R Environment. E.2 R Code for Nonparametric Methods. E.3 R Code for Parametric and Mixed Methods. E.4 R Code for the Study of Uncertainty. E.5 Other R Functions. References. Index.

Evaluating the Foreign Ownership Wage Premium Using a Difference-in-Differences Matching Approach

Abstract: This paper seeks to identify the causal effect of foreign acquisitions on wages of skilled and unskilled workers, using difference-in-differences propensity score matching estimators. Our results suggest that there is substantial heterogeneity in the post-acquisition wage effect depending on the nationality of the foreign acquirer and the skill group of workers. We find sizable post acquisition wage effects on skilled and unskilled wages following an acquisition by a US firm. No such impacts result from acquisitions by EU multinationals. Also we discern some positive wage effects for unskilled workers resulting from acquisitions by multinationals from the rest of the world.

The Shazam music recognition service

Abstract: Guided by a user's query-by-example music sample, it delivers the matching song, as well as related music information, of immediate interest to the user.

Frictional Wage Dispersion in Search Models : A Quantitative Assessment, Working Paper 06-07

Abstract: Standard search and matching models of equilibrium unemployment, once properly calibrated, can generate only a small amount of frictional wage dispersion, i.e., wage differentials among ex-ante similar workers induced purely by search frictions. We derive this result for a specific measure of wage dispersion

Contour map matching for event detection in sensor networks

Abstract: Many sensor network applications, such as object tracking and disaster monitoring, require effective techniques for event detection. In this paper, we propose a novel event detection mechanism based on matching the contour maps of in-network sensory data distribution. Our key observation is that events in sensor networks can be abstracted into spatio-temporal patterns of sensory data and that pattern matching can be done efficiently through contour map matching. Therefore, we propose simple SQL extensions to allow users to specify common types of events as patterns in contour maps and study energy-efficient techniques of contour map construction and maintenance for our pattern-based event detection. Our experiments with synthetic workloads derived from a real-world coal mine surveillance application validate the effectiveness and efficiency of our approach.

Beauty is a Beast, Frog is a Prince: Assortative Matching with Nontransferabilities

Abstract: We present sufficient conditions for monotone matching in environments where utility is not fully transferable between partners. These conditions involve not only complementarity in types of the total payoff to a match, as in the transferable utility case, but also monotonicity in type of the degree of transferability between partners. We apply our conditions to study some models of risk sharing and incentive problems, deriving new results for predicted matching patterns in those contexts.

An Assessment of Propensity Score Matching as a Nonexperimental Impact Estimator Evidence from Mexico's PROGRESA Program

Abstract: In this working paper the authors present evidence on the reliability of propensity score matching to estimate the bias associated with the effect of treatment on the treated, exploiting the availability of experimental data from a Mexican antipoverty program (PROGRESA: Programa de Educacion, Salud y Alimentacion). The data comes from several outcomes such as food expenditure and child schooling and labor. The methodology compares the results of the experimental impact estimator with those using matched samples drawn from a (non-experimental) national survey carried out to measure household income and expenditures. The results show that simple-cross sectional matching produces significant bias for outcomes measured in different ways. Results are more positive for outcomes measured similarly across survey instruments, but even in this case there are indications of bias depending on sample and matching method.

Matching Structure and Semantics: A Survey on Graph-Based Pattern Matching.

Abstract: The task of matching patterns in graph-structured data has applications in such diverse areas as computer vision, biology, electronics, computer aided design, social networks, and intelligence analysis. Consequently, work on graph-based pattern matching spans a wide range of research communities. Due to variations in graph characteristics and application requirements, graph matching is not a single problem, but a set of related problems. This paper presents a survey of existing work on graph matching, describing variations among problems, general and specific solution approaches, evaluation techniques, and directions for further research. An emphasis is given to techniques that apply to general graphs with semantic characteristics.

Image Matching by Normalized Cross-Correlation

Abstract: Correlation is widely used as an effective similarity measure in matching tasks. However, traditional correlation based matching methods are limited to the short baseline case. In this paper we propose a new correlation based method for matching two images with large camera motion. Our method is based on the rotation and scale invariant normalized cross-correlation. Both the size and the orientation of the correlation windows are determined according to the characteristic scale and the dominant direction of the interest points. Experimental results on real images demonstrate that the new method is effective for matching image pairs with significant rotation and scale changes as well as other common imaging conditions.

Survey on Block Matching Motion Estimation Algorithms and Architectures with New Results

Abstract: Block matching motion estimation is the heart of video coding systems. During the last two decades, hundreds of fast algorithms and VLSI architectures have been proposed. In this paper, we try to provide an extensive exploration of motion estimation with our new developments. The main concepts of fast algorithms can be classified into six categories: reduction in search positions, simplification of matching criterion, bitwidth reduction, predictive search, hierarchical search, and fast full search. Comparisons of various algorithms in terms of video quality and computational complexity are given as useful guidelines for software applications. As for hardware implementations, full search architectures derived from systolic mapping are first introduced. The systolic arrays can be divided into inter-type and intra-type with 1-D, 2-D, and tree structures. Hexagonal plots are presented for system designers to clearly evaluate the architectures in six aspects including gate count, required frequency, hard-ware utilization, memory bandwidth, memory bitwidth, and latency. Next, architectures supporting fast algorithms are also reviewed. Finally, we propose our algorithmic and architectural co-development. The main idea is quick checking of the entire search range with simplified matching criterion to globally eliminate impossible candidates, followed by finer selection among potential best matched candidates. The operations of the two stages are mapped to the same hardware for resource sharing. Simulation results show that our design is ten times more area-speed efficient than full search architectures while the video quality is competitively the same.

Do Subsidies Increase Charitable Giving in the Long Run? Matching Donations in a Field Experiment

Abstract: Subsidizing charitable giving, for example, for victims of natural disasters, is very popular, not only with governments but also with private organizations. Many companies, for example, match their employees' charitable contributions, hoping that this will foster the willingness to contribute. However, systematic analyses of the effect of such a matching mechanism are still lacking. This paper tests the effect of matching charitable giving in a randomized field experiment in the short and the long run. The donations of a randomly selected group were matched by contributions from an anonymous donor. The results support the hypothesis that a matching mechanism increases contributions to a public good. However, in the periods after the experiment, when matching donations have been stopped, the contribution rate declines for the treatment group. The matching mechanism leads to a negative net effect on the participation rate. The field experiment therefore provides evidence suggesting that the willingness to contribute may be undermined by a matching mechanism in the long run.

Semantic WS-agreement partner selection

Abstract: In a dynamic service oriented environment it is desirable for service consumers and providers to offer and obtain guarantees regarding their capabilities and requirements. WS-Agreement defines a language and protocol for establishing agreements between two parties. The agreements are complex and expressive to the extent that the manual matching of these agreements would be expensive both in time and resources. It is essential to develop a method for matching agreements automatically. This work presents the framework and implementation of an innovative tool for the matching providers and consumers based on WS-Agreements. The approach utilizes Semantic Web technologies to achieve rich and accurate matches. A key feature is the novel and flexible approach for achieving user personalized matches.

Dynamic resource matching system

Abstract: A method, apparatus, and system are provided for implementing resource and/or location-based matching services between a wireless terminal (e.g., mobile phone) user and one or more resources. A novel infrastructure supports resource and/or location based matching services over a wireless network. A back-end system includes a database, server, and match engine that are configured match a user with one or more resources based on the user's characteristics, preferences, and/or location. Such resources include (1) other users, (2) targeted advertising, (3) businesses/networking opportunities, and/or (4) locate a nearby service or store. A flexible database architecture supports application-specific resources which facilitate the deployment of various matching services. Application developers are thus able to implement different resource-matching applications for wireless devices through a common back-end infrastructure. Additionally, the match engine may include a feedback mechanism that permits the match engine to learn a user's preferences.