An understanding of what influences sinkhole formation and the ability to accurately predict sinkhole hazards is critical to environmental management efforts in the karst lands of southeastern Minnesota. Based on the distribution of distances to the nearest sinkhole, sinkhole density, bedrock geology and depth to bedrock in southeastern Minnesota and northwestern Iowa, a decision tree model has been developed to construct maps of sinkhole probability in Minnesota. The decision tree model was converted as cartographic models and implemented in ArcGIS to create a preliminary sinkhole probability map in Goodhue, Wabasha, Olmsted, Fillmore, and Mower Counties. This model quantifies bedrock geology, depth to bedrock, sinkhole density, and neighborhood effects in southeastern Minnesota but excludes potential controlling factors such as structural control, topographic settings, human activities and land-use. The sinkhole probability map needs to be verified and updated as more sinkholes are mapped and more information about sinkhole formation is obtained.

No description available.

L'expansion du média Internet pour le recrutement a entraîné ces dernières années la multiplication des canaux dédiés à la diffusion des offres d'emploi. Dans un contexte économique où le contrôle des coûts est primordial, évaluer et comparer les performances des différents canaux de recrutement est devenu un besoin pour les entreprises. Cette thèse a pour objectif le développement d'un outil d'aide à la décision destiné à accompagner les recruteurs durant le processus de diffusion d'une annonce. Il fournit au recruteur la performance attendue sur les sites d'emploi pour un poste à pourvoir donné. Après avoir identifié les facteurs explicatifs potentiels de la performance d'une campagne de recrutement, nous appliquons aux annonces des techniques de fouille de textes afin de les structurer et d'en extraire de l'information pertinente pour enrichir leur description au sein d'un modèle explicatif. Nous proposons dans un second temps un algorithme prédictif de la performance des offres d'emploi, basé sur un système hybride de recommandation, adapté à la problématique de démarrage à froid. Ce système, basé sur une mesure de similarité supervisée, montre des résultats supérieurs à ceux obtenus avec des approches classiques de modélisation multivariée. Nos expérimentations sont menées sur un jeu de données réelles, issues d'une base de données d'annonces publiées sur des sites d'emploi. / Last years, e-recruitment expansion has led to the multiplication of web channels dedicated to job postings. In an economic context where cost control is fundamental, assessment and comparison of recruitment channel performances have become necessary. The purpose of this work is to develop a decision-making tool intended to guide recruiters while they are posting a job on the Internet. This tool provides to recruiters the expected performance on job boards for a given job offer. First, we identify the potential predictors of a recruiting campaign performance. Then, we apply text mining techniques to the job offer texts in order to structure postings and to extract information relevant to improve their description in a predictive model. The job offer performance predictive algorithm is based on a hybrid recommender system, suitable to the cold-start problem. The hybrid system, based on a supervised similarity measure, outperforms standard multivariate models. Our experiments are led on a real dataset, coming from a job posting database.

L'objectif de cette thèse est de fournir des mécanismes améliorés pour déployer des services compétitifs par des manières rapides et rentables. Nous proposons un modèle de composition de service basé sur un environnement IMS/Web convergent. Ce modèle permet aux utilisateurs non professionnels de réutiliser les services existants pour créer de nouveaux services facilement. Pour améliorer la fonctionnalité de composition automatique, trois stratégies, y compris mise à jour passive, mise à jour active et mise à jour hybride sont proposées et analysées. Nous introduisons ensuite une plateforme centralisée d'exposition de service pour une variété de services, y compris services de Télécom / Web / appareil / services générés par les utilisateurs. Cette plateforme vise à renforcer les caractéristiques de centrée-sur-utilisateur et convergence, et fournir l'accès unifié à différents services. Par la suite, deux modèles basés sur le P2P sont conçus pour compléter le modèle centralisé: i) Un modèle hiérarchique basé sur Chord pour garantir l'efficacité de la découverte de services. Il adopte le concept de publication et découverte de service abstrait pour permettre à la recherche de service ambiguë. ii) Un modèle de superposition-triplex et P2P basé, qui cible principalement des services offerts par les appareils. Dans ce modèle, nous utilisons des passerelles pour déléguer des appareils résidant en eux pour l'exposition globale de services, et utilisons une architecture basée sur une superposition triplex, qui comprend une couche P2P non structurée, une couche de réseau sémantique (SON), et une couche de dépendance de service, pour la partage de l'information de service et la découverte de services / The goal of this thesis is to provide enhanced mechanisms to deploy competitive services in a rapid and cost-effective manner. To achieve this goal, we first propose an automatic service composition model relying on an IMS/Web converged environment. This service composition model is intended to be one in which even non-professional users can easily reuse existing services to create new services. To further improve the automatic service composition feature, three strategies including passive update, active update and hybrid update are proposed and analyzed. We then propose a centralized service exposure framework for a variey of services, including Telecom / Web / Device / user-generated services. This framework aims at enhancing the user-centric and convergence features, and providing the unified access to diverse services.Subsequently, two P2P based service information sharing models are designed to complement the centralized service exposure model : i) A hierarchical P2P based model, which reuses Chord for guaranteeing the service discovery efficiency, meanwhile adopts the concept of abstract service publication and discovery for enabling the ambiguous services searching. ii) A triplex P2P overlay based model, which mainly targets the devices offered services. In this model, we use the gateways to delegate the devices residing in them for the global service exposure, and use a triplex overlay based architecture, which includes an underlying unstructured P2P layer, a Semantic Overlay Network (SON) based overlay and a service dependency overlay, for the service information sharing and discovery

Un nombre croissant de sources de données interconnectées sont publiées sur le Web. Cependant, leur schéma peut êtreincomplet ou absent. De plus, les données ne sont pas nécessairement conformes au schéma déclaré. Ce qui rend leur exploitation complexe. Dans cette thèse, nous proposons une approche d’extraction automatique et incrémentale du schéma d’une source à partir de la structure implicite de ses données. Afin decompléter la description des types découverts, nous proposons également une approche de découverte des patterns structurels d’un type. L’approche procède en ligne sans avoir à télécharger ou à parcourir la source. Ce quipeut être coûteux voire impossible car les sources sont interrogées à distance et peuvent imposer des contraintes d’accès, notamment en termes de temps ou de nombre de requêtes. Nous avons abordé le problème de l’annotationafin de trouver pour chaque type un ensemble de labels permettant de rendre compte de sonsens. Nous avons proposé des algorithmes d’annotation qui retrouvent le sens d’un type en utilisant des sources de données de références. Cette approche s’applique aussi bien pour trouver des noms pertinents pour les typesdécouverts que pour enrichir la description des types existants. Enfin, nous nous sommes intéressés à caractériser la conformité entre les données d’une source et le schéma qui les décrit. Nous avons proposé une approche pour l'analyse et l'amélioration de cette conformité et nous avons proposé des facteurs de qualité, les métriques associées, ainsi qu'une extension du schéma permettant de refléter l'hétérogénéité entre les instances d'un type. / An increasing number of linked data sources are published on the Web. However, their schema may be incomplete or missing. In addition, data do not necessarily follow their schema. This flexibility for describing the data eases their evolution, but makes their exploitation more complex. In our work, we have proposed an automatic and incremental approach enabling schema discovery from the implicit structure of the data. To complement the description of the types in a schema, we have also proposed an approach for finding the possible versions (patterns) for each of them. It proceeds online without having to download or browse the source. This can be expensive or even impossible because the sources may have some access limitations, either on the query execution time, or on the number of queries.We have also addressed the problem of annotating the types in a schema, which consists in finding a set of labels capturing their meaning. We have proposed annotation algorithms which provide meaningful labels using external knowledge bases. Our approach can be used to find meaningful type labels during schema discovery, and also to enrichthe description of existing types.Finally, we have proposed an approach to evaluate the gap between a data source and itsschema. To this end, we have proposed a setof quality factors and the associated metrics, aswell as a schema extension allowing to reflect the heterogeneity among instances of the sametype. Both factors and schema extension are used to analyze and improve the conformity between a schema and the instances it describes

We are living within the age of information. The ever increasing flow of data and publications poses a monumental bottleneck to scientific progress as despite the amazing abilities of the human mind, it is woefully inadequate in processing such a vast quantity of multidimensional information. The small bits of flotsam and jetsam that we leverage belies the amount of useful information beneath the surface. It is imperative that automated tools exist to better search, retrieve, and summarize this content. Combinations of document indexing and search engines can quickly find you a document whose content best matches your query - if the information is all contained within a single document. But it doesn’t draw connections, make hypotheses, or find knowledge hidden across multiple documents. Literature-based discovery is an approach that can uncover hidden interrelationships between topics by extracting information from existing published scientific literature. The proposed study utilizes a semantic-based approach that builds a graph of related concepts between two user specified sets of topics using semantic predications. In addition, the study includes properties of bibliographically related documents and statistical properties of concepts to further enhance the quality of the proposed intermediate terms. Our results show an improvement in precision-recall when incorporating citations.

Indiana University-Purdue University Indianapolis (IUPUI) / Few writing teachers will disagree that teaching writing conventions in isolation is a fruitless, even harmful, pedagogy which does little, if anything, to improve student writing. Teaching conventions, style, and usage (often collectively referred to as grammar) in context, however, proves difficult when struggling secondary students develop good ideas and evidence but fail to clearly articulate them because of their lack of understanding of various writing conventions. The purpose of this study is to test the efficacy of a carefully designed discovery learning activity which intends to push students into metacognition about what they read, how it is structured, and how that structure affects the reader. Three sources of data were used to determine whether students who had learned by discovery were better able to avoid and revise run-on sentences than students who did not learn through discovery pedagogy. The data sources include two sets of essays, surveys taken by the students, and teacher analyses of essays for readability. The results of the data analysis indicate that use of run-on sentences, especially early in an essay, detrimentally affects the readability of student written work; discovery learning activities improve student understanding, application, and transfer of skill; and while students believe they understand more than their written work indicates, the results provide teachers direction for further instruction. The findings of this study indicate that use of discovery learning for writing instruction with struggling learners holds great promise: a group of students generally regarded as academically weak showed greater understanding and application of run-on sentence avoidance than slightly stronger students who learned without discovery methods. This indicates that discovery learning is a method that improves learning among reluctant secondary students, a population many teachers struggle to reach effectively. Discovery learning is not limited to conventions, though: the promise of its application potential extends into a variety of writing skills and concepts. In addition to the run-on sentence discovery activity studied here, discovery activities for various other skills—from semicolon use through creating characterization with dialogue—are included.

This master's thesis deals with questions of the data mining and an extension of a data mining system in the Oracle environment developed at FIT. So far, this system cannot apply to real-life conditions as there are no data mining modules available. This system's core application design includes an interface allowing the addition of mining modules. Until now, this interface has been tested on a sample mining module only; this module has not been executing any activity just demonstrating the use of this interface. The main focus of this thesis is the study of this interface and the implementation of a functional mining module testing the applicability of the implemented interface. Association rule mining module was selected for implementation.

No description available.

No description available.