Image matching is an important and necessary process in Content-Based Image Retrieval (CBIR). We propose three new methods for image matching: the first one is based on the Local Triplet Pattern (LTP) histograms; the second one is based on the Gaussian Mixture Models (GMMs) estimated by using the Extended Mass-constraint (EMass) algorithm; and the third one is called the DCT2KL algorithm. First, the LTP histograms are proposed to capture spatial relationships between color levels of neighboring pixels. An LTP level is extracted from each 3x3 pixel block, which is a unique number describing the color level relationship between a pixel and its neighboring pixels. Second, we consider how to represent and compare multi-dimensional color features using GMMs. GMMs are alternative methods to histograms for representing data distributions. GMMs address the high-dimensional problems from which histograms usually suffer inefficiency. In order to avoid local maxima problems in most GMM estimation algorithms, we apply the deterministic annealing method to estimate GMMs. Third, motivated by image compression algorithms, the DCT2KL method addresses the high dimensional data by using the Discrete Cosine Transform (DCT) coefficients in the YCbCr color space. The DCT coefficients are restored by partially decoding JPEG images. Assume that each DCT coefficient sequence is emitted from a memoryless source, and all these sources are independent of each other. For each target image we form a hypothesis that its DCT coefficient sequences are emitted from the same sources as the corresponding sequences in the query image. Testing these hypotheses by measuring the log-likelihoods leads to a simple yet efficient scheme that ranks each target image according to the Kullback-Leibler (KL) divergence between the empirical distribution of the DCT coefficient sequences in the query image and that in the target image. Finally we present a scheme to combine different features and methods to boost the performance of image retrieval. Experimental results on different image data sets show that these three methods proposed above outperform the related works in literature, and the combination scheme further improves the retrieval performance.

This thesis examines determinants of the adoption of total mixed ration (TMR), and the effects of the adoption of TMR on the farm level productivity and profitability of Ontario dairy farms. A sample of 320 farm level data from 2004-2008 is taken from the Ontario Dairy Farm Accounting Project (ODFAP). A probit model is estimated to examine the factors affecting the adoption of TMR; and the propensity score matching analysis is used to explore the influence of the use of TMR on sample farm’s productivity and profitability. Results from the probit model show that farmer’s age, herd size, region, breed type and feeding system have significant effect on the adoption of TMR. In turn, the adoption of TMR feeding practice has positive influence on both farm productivity and profitability. Under the propensity score matching method, the use of the TMR feeding practice has an economically significant effect on farm profits (i.e., for average farm with approximately 73 cows, the use of TMR feeding practice increases farm profits by CAD$37,091.30/year approximately) and a statistically significant increase in milk production by 1075.41 hl/cow per year.

In this thesis, we study a variant of the maximum cardinality matching problem known as the maximum charge problem. Given a graph with arbitrary positive integer capacities assigned on every vertex and every edge, the goal is to maximize the assignment of positive feasible charges on the edges obeying the capacity constraints, so as to maximize the total sum of the charges. We use the primal-dual approach. We propose a combinatorial algorithm for solving the dual of the restricted primal and show that the primal-dual algorithm runs in a polynomial time. / ix, 96 leaves : ill. ; 29 cm

This thesis uses a macroeconomic approach to study labour adjustments following sector-specific shocks. I develop a general model, investigate its dynamic adjustment process and apply it to study the Canadian economy in 2002–2006. This episode is an interesting case study because it features a significant labour reallocation to the resource sector and away from manufacturing, precipitated by an increase in global commodity prices and an associated exchange rate appreciation. The results establish that impediments to the adjustment process are economically significant in the aggregate for this episode, imposing costs of up to three percent of output during the transition. These findings augment several studies that suggest individual workers can face large and persistent earnings losses during job turnover. However, unlike previous research, I use the search and matching approach — which incorporates explicit labour market frictions — to uncover the sources of these costs for the macroeconomy. The findings emphasize that job loss itself is not particularly important quantitatively, but rather the non-transferability of skills during job turnover is a key concern. Finally, I investigate how labour market policy impacts the economy’s response to sector-specific shocks by analyzing a counterfactual policy change in unemployment benefits and improved skill acquisition through faster learning and training subsidies. The results reveal interesting policy trade-offs. First, I find that increasing unemployment benefits prolongs the economy’s adjustment, reduces employment, output and welfare and increases unemployment incidence and duration. However, because this policy impacts high-productivity and low-productivity sectors differently, it shifts the composition of the remaining jobs towards high-productivity sectors, thereby raising aggregate productivity and also reduces wage inequality. Second, I find that faster skill acquisition has the potential to deliver large economic gains in the long-run, but requires up-front investment costs which entail reduced economic performance in the short-run. / Thesis (Ph.D, Economics) -- Queen's University, 2008-08-05 23:44:39.827

No description available.

No description available.

This study investigated a local aspect of theSwedish elementary school modernization process between the years 1871-1882.Previous studies have been drawn from the school inspectors own reports andshow that the schoolboards housed a negative and resentful attitude towardsreforms and the school inspectors often felt that the parish priests were moreinclined to changes than the schoolboards themselves.  Other research on elementary schoolmodernization has stressed the matching grants as a major cause of developmentand further establishment. In this study, however, a schoolboards own notes inthe form of meeting protocols during an eleven year period as well as a letterby a priest Erik Lundberg, Redogörelseför skolhusfrågan i Tierp dated 6 October 1880 were analyzed to see if theprevious explanation of school modernization, in particular with regard to thereluctance to change by the schoolboards goes to demonstrate and confirm. Theresults of this study have shown that even if the schoolboards do notimmediately follow the school inspectors instructions, it seems not primarilybeen due to resentment. Reforms within the school world are well known forbeing time-consuming but for the current period, as this study concerns andwithin this specific parish there were conflicts of interest which are notshown when only the inspector’s accounts and reports are examined. The late 1900thcentury was a time with a great deal of overwhelming and overthrowingreformations and changes witch challenged its previous social structuresthrough industrialization, urbanization and new political currents. Developmentand change occurs faster than news of it could travel, and it has been aninteresting journey to study these documents and to get a glimpse into how aschoolboard in a rural parish dealt with expansion and change.

We compare two popular scenario tree generation methods in the context of financial optimization: Moment matching and scenario reduction. Using a simple problem with a known analytic solution, we find that moment matching - accompanied by a check to ensure absence of arbitrage opportunities - replicates this solution precisely. On the other hand, even if the scenario trees generated by scenario reduction are arbitrage-free, the solutions to the approximate optimization problem represented by the reduced tree are biased and highly variable. These results hold for correlated and uncorrelated asset returns, as well as for normal and non-normal returns. (authors' abstract)

Research suggests that individuals with severe developmental disabilities are able to indicate preferences during direct preference assessments with objects. However, they may not have the discrimination skills to respond effectively to pictures. Recent research has begun to focus on identifying and teaching individuals the skills needed to indicate their preferences using pictures, due to the practical advantages of presenting choices with pictures versus objects. Further, it has been suggested that generalized identity matching may be associated with successful performance during picture preference assessments. The present study evaluated whether teaching generalized identity matching with pictures would lead to improved performance on picture preference assessments. Two adults, one male and one female, diagnosed with moderate to severe mental retardation, with limited to no speech, and varying degrees of physical impairments participated in the study. Before training, participants were able to indicate their preferences using objects but not with pictures and they were unable to perform generalized identity matching. Participants were trained on an object-picture matching-to-sample relation in a multiple baseline across tasks design. Individual data showed that neither Participant 1 nor Participant 2 met the pass criterion (80% or higher correct responses) on at least one generalized identity matching task after one training task was taught. Further, both participants showed improved concordance between preference assessments using objects and pictures of the same objects after mastering one training task respectively. These findings were independent of the accuracy of responding during the generalized identity matching post-assessment for each participant. These results suggest that the ability to respond accurately on generalized picture-picture identity matching tasks may not be related to performance on picture preference assessments (i.e., for individuals to indicate their preferences with pictures).

This thesis focuses on three topics relevant to humanitarian applications: (i) stable and complete assignment of staff members to field offices, (ii) bottleneck management for transportation networks, and (iii) performance measurement of the food assistance supply chain. The assignment and reassignment of personnel to jobs is a large-scale problem faced by many organizations including the military and multi-national organizations. Although successful algorithms have been developed that can ensure matchings that are stable (without incentive to deviate), not all practical concerns have been addressed by these algorithms. For example, the gap we study is that when staff members do not provide preference lists covering all jobs, a complete stable matching is not guaranteed. In the first part of the thesis, we model negotiations, which occur in practice, as part of the problem of matching all agents. We introduce algorithms and structural results for when the organization negotiates with specific agents to modify their preference lists and the centralized objective is to minimize the number or cost of negotiations required to achieve complete stable matchings. An uncertain environment with disruptions is a reality faced by many humanitarian operations but not fully addressed in the literature. Transportation delays are often driven by reliability issues (e.g., customs delays, strikes, and the availability of transport), and the length of wait time can be influenced by congestion. In the second part of the thesis, we describe a queuing model with breakdowns to model delays in port and transportation corridors (the overland travel from discharge ports to delivery points). Using the model, we gain insights into where delays are most detrimental to system performance (i.e., the network's "bottleneck") in port and transportation corridors. We then include our delay modeling in a convex cost network flow model that determines optimal routing when several port and corridor options are available. Finally, we examine a resource allocation model for where to invest in improvements to minimize delay. Throughout, we compare solutions using the optimal approach to rules of thumb and identify important factors that might be missing in practical decision making currently. Third, we present a case study on the implementation of supply chain key performance indicators (KPIs) at a large humanitarian organization. We describe (i) the phases necessary for a full implementation of supply chain KPIs at a humanitarian or non-profit organization, (ii) how to address strategy, mindset, and organizational barriers, and (iii) how to adapt commercial supply chain KPI frameworks to the humanitarian sector, factoring in implementation constraints present in the humanitarian sector that may impact KPI development. Last, a conclusion chapter discusses areas where this research may or may not generalize for each of the three topics studied.