In order to address the growing waste concerns facing South Africa, the Department of Environmental Affairs (DEA) published the National Waste Management Strategy in 2012. The ultimate aim of the strategy is to effect the realisation of the objectives of the National Environmental Management Waste Act (NEMWA). The NEMWA is a sector environmental law based on the waste management hierarchy and its provisions apply to waste management activities in all three government spheres amongst others. The NWMS provides for eight distinct goals with accompanying targets to be met by 2016. Some of these goals and targets speak directly to the solid waste management mandate of local government. Concerns about the capacity of municipalities to see the execution of this mandate however, raise questions about the likelihood of some NWMS goals and targets being met by 2016. The North West Province serves as a case in point where recent reports by the office of the Auditor General have highlighted significant non-compliance in local government with the provisions of the existing waste law and policy framework of South Africa. It appears form these reports that a number of performance gaps and capacity constraints exist as far as it concerns municipal solid waste management. Against this back ground the study questions the governance strategies necessary within the municipalities of the North West Province to: a) address the existing performance gaps and capacity constraints and b) progressively move towards meeting the NWMS goals and targets as far as it concerns solid waste management. The study will explore the objectives, goals and targets of the NWMS against the background of the National Environmental Management Act (NEMA) and NEMWA, with a specific focus on solid waste management. The study will further review the documented performance gaps and capacity constraints as a far as it concerns solid waste management in municipalities in the North West province, specifically. / M. Environmental Management, North-West University, Potchefstroom Campus, 2014

In order to address the growing waste concerns facing South Africa, the Department of Environmental Affairs (DEA) published the National Waste Management Strategy in 2012. The ultimate aim of the strategy is to effect the realisation of the objectives of the National Environmental Management Waste Act (NEMWA). The NEMWA is a sector environmental law based on the waste management hierarchy and its provisions apply to waste management activities in all three government spheres amongst others. The NWMS provides for eight distinct goals with accompanying targets to be met by 2016. Some of these goals and targets speak directly to the solid waste management mandate of local government. Concerns about the capacity of municipalities to see the execution of this mandate however, raise questions about the likelihood of some NWMS goals and targets being met by 2016. The North West Province serves as a case in point where recent reports by the office of the Auditor General have highlighted significant non-compliance in local government with the provisions of the existing waste law and policy framework of South Africa. It appears form these reports that a number of performance gaps and capacity constraints exist as far as it concerns municipal solid waste management. Against this back ground the study questions the governance strategies necessary within the municipalities of the North West Province to: a) address the existing performance gaps and capacity constraints and b) progressively move towards meeting the NWMS goals and targets as far as it concerns solid waste management. The study will explore the objectives, goals and targets of the NWMS against the background of the National Environmental Management Act (NEMA) and NEMWA, with a specific focus on solid waste management. The study will further review the documented performance gaps and capacity constraints as a far as it concerns solid waste management in municipalities in the North West province, specifically. / M. Environmental Management, North-West University, Potchefstroom Campus, 2014

Les Interfaces Cerveau-Machine sont des dispositifs permettant d'instaurer un canal de communication entre le cerveau humain et le monde extérieur sans utiliser les voies usuelles nerveuses et musculaires. Le développement de tels systèmes se situe à l'interface entre le traitement du signal, l'apprentissage statistique et la neurophysiologie. Dans cette thèse, nous avons réalisé et étudié un dispositif d'Interface Cerveau-Machine non invasif asynchrone, c'est-à-dire capable d'identifier des actions mentales associées à des tâches motrices ou cognitives imaginées sans synchronisation sur un événement contrôlé par un système externe. Celui-ci est basé sur l'analyse en temps réel de signaux électro-encéphalographiques (EEG) issus d'électrodes disposées à la surface de la tête d'un sujet humain. Du point de vue méthodologique, nous avons implémenté plusieurs techniques de prétraitement de ces signaux et comparé leur influence sur les performances du système. Ces techniques comprennent : 1) l'utilisation directe du signal issu des capteurs EEG, 2) l'exploitation de méthodes de séparation de sources qui permettent de résumer les signaux EEG par un faible nombre de composantes spatiales et 3) la reconstruction de l'activité des sources de courant corticales par résolution du problème inverse en EEG. De plus, plusieurs mesures permettant de quantifier l'activité cérébrale sont exploitées et comparées : la puissance spectrale, la cohérence et la synchronie de phase. Nos résultats montrent que la reconstruction préalable de l'activité corticale par problème inverse, ainsi que l'utilisation de mesures d'interaction à distance permettent d'améliorer les performances du dispositif.

According to the Sectional Titles Act 95 of 1986, levies are assigned according to the size of a unit, in other words a unit’s participation quota. These levies are used to finance a complex’s insurance, common property electricity and water, lift maintenance, audit fees, management agent fees, salaries and wages, security, swimming pool expenses and general building maintenance. According to the Traditional Costing System, indirect costs are allocated based on a single cost actual expenses over a year are dissected. A regression and correlation analysis was done on the relationship between costs, participation quota and levies. The contribution of the study is that it empirically determines the behaviour of cost items in order to allocate indirect costs more accurately. Consequently, this will result in improved reasonability and will eliminate the cross-subsidisation of units. The objectives were reached as follows: The first objective was reached when it was proved through empirical studies that the Sectional Title is not fair when it states that the allocation of levies must be according to floor space. The second objective was reached when both participation quota and units per complex were identified as cost drivers for each cost item. The third objective was reached by allocating the levy on a 50 percent base according to units and a 50 percent base according to the total area. This could be refined by each complex according to the unique need of the complex. driver. However, various indirect costs are not necessarily subject to this single cost driver. In practice, it has been proven that high volume products are proportionally taxed with indirect costs when compared to low volume products. In many complexes, owners with larger units are of the opinion that their larger units are subsidising those owners with smaller units. In contrast to the traditional costing system, activity-based costing recognises that indirect costs can also be assigned by use of multiple cost drivers. Since these multiple cost drivers – which affect the way costs are assigned – can be identified, indirect costs are assigned more accurately. This study investigates the current method of cost allocation, whereby costs are assigned via participation quota, and makes suggestions on how these costs can be assigned on a more accurate and fair basis in practice. The goal of this study is to challenge the reasonability of the Sectional Titles Act. The objectives of the dissertation are firstly, to determine the cost behaviour of various cost items and whether the participation quota is a fair cost driver for determining the levies that are to be paid; secondly, it investigates alternative cost drivers that will be more applicable to certain cost items; thirdly, to suggest a cost formula to replace the current cost method, that being the allocation of indirect costs via participation quota. Empirical methods have been used in the research. The empirical research was performed using data obtained from ANGOR Property Specialists (Pty) Ltd’s database. A sample of 113 complexes was extracted from the database of which the / Thesis (M.Com. (Management Accountancy))--North-West University, Potchefstroom Campus, 2011

Cette thèse propose une méthodologie pour des études en simulation numérique en grande dimension. Elle se décline en différentes étapes : construction de plan d'expériences approprié, analyse de sensibilité et modélisation par surface de réponse. Les plans d'expériences adaptés à la simulation numérique sont les "Space Filling Designs", qui visent à répartir uniformément les points dans l'espace des variables d'entrée. Nous proposons l'algorithme WSP pour construire ces plans, rapidement, avec de bons critères d'uniformité, même en grande dimension. Ces travaux proposent la construction d'un plan polyvalent, qui sera utilisé pour les différentes étapes de l'étude : de l'analyse de sensibilité aux surfaces de réponse. L'analyse de sensibilité sera réalisée avec une approche innovante sur les points de ce plan, pour détecter le sous-ensemble de variables d'entrée réellement influentes. Basée sur le principe de la méthode de Morris, cette approche permet de hiérarchiser les variables d'entrée selon leurs effets. Le plan initial est ensuite "replié" dans le sous-espace des variables d'entrée les plus influentes, ce qui nécessite au préalable une étude pour vérifier l'uniformité de la répartition des points dans l'espace réduit et ainsi détecter d'éventuels amas et/ou lacunes. Ainsi, après réparation, ce plan est utilisé pour l'étape ultime : étude de surfaces de réponse. Nous avons alors choisi d'utiliser l'approche des Support Vector Regression, indépendante de la dimension et rapide dans sa mise en place. Obtenant des résultats comparables à l'approche classique (Krigeage), cette technique semble prometteuse pour étudier des phénomènes complexes en grande dimension. / This thesis proposes a methodology of study in numeric simulation for high dimensions. There are several steps in this methodology : setting up an experimental design, performing sensitivity analysis, then using response surface for modelling. In numeric simulation, we use a Space Filling Design that scatters the points in the entire domain. The construction of an experimental design in high dimensions must be efficient, with good uniformity properties. Moreover, this construction must be fast. We propose using the WSP algorithm to construct such an experimental design. This design is then used in all steps of the methodology, making it a versatile design, from sensitivity analysis to modelling. A sensitivity analysis allows identifying the influent factors. Adapting the Morris method principle, this approach classifies the inputs into three groups according to their effects. Then, the experimental design is folded over in the subspace of the influent inputs. This action can modify the uniformity properties of the experimental design by creating possible gaps and clusters. So, it is necessary to repair it by removing clusters and filling gaps. We propose a step-by-step approach to offer suitable repairing for each experimental design. Then, the repaired design is used for the final step: modelling from the response surface. We consider a Support Vector Machines method because dimension does not affect the construction. Easy to construct and with good results, similar to the results obtained by Kriging, the Support Vector Regression method is an alternative method for the study of complex phenomena in high dimensions.

Le cancer de la prostate est la deuxième cause de mortalité chez l’homme en France. L’IRM multiparamétrique est considérée comme la technique la plus prometteuse pour permettre une cartographie du cancer, ouvrant la voie au traitement focal, alternatif à la prostatectomie radicale. Néanmoins, elle reste difficile à interpréter et est sujette à une forte variabilité inter- et intra-expert, d’où la nécessité de développer des systèmes experts capables d’aider le radiologue dans son diagnostic. Nous proposons un système original d’aide au diagnostic (CAD) offrant un second avis au radiologue sur des zones suspectes pointées sur l’image. Nous évaluons notre système en nous appuyant sur une base de données clinique de 30 patients, annotées de manière fiable et exhaustive grâce à l’analyse des coupes histologiques obtenues par prostatectomie. Les performances mesurées dans des conditions cliniques auprès de 12 radiologues, sans et avec notre outil, démontrent l’apport significatif de ce CAD sur la qualité du diagnostic, la confiance des radiologues et la variabilité inter-expert. La création d’une base de corrélations anatomo-radiologiques est une tâche complexe et fastidieuse. Beaucoup d’études n’ont pas d’autre choix que de s’appuyer sur l’analyse subjective d’un radiologue expert, entâchée d’incertitude. Nous proposons un nouveau schéma de classification, basé sur l’algorithme du séparateur à vaste marge (SVM), capable d’intégrer, dans la fonction d’apprentissage, l’incertitude sur l’appartenance à une classe (ex. sain/malin) de certains échantillons de la base d’entraînement. Les résultats obtenus, tant sur des exemples simulés que sur notre base de données cliniques, démontrent le potentiel de ce nouvel algorithme, en particulier pour les applications CAD, mais aussi de manière plus générale pour toute application de machine learning s’appuyant sur un étiquetage quantitatif des données / Prostate cancer is one of the leading cause of death in France. Multi-parametric MRI is considered the most promising technique for cancer visualisation, opening the way to focal treatments as an alternative to prostatectomy. Nevertheless, its interpretation remains difficult and subject to inter- and intra-observer variability, which motivates the development of expert systems to assist radiologists in making their diagnosis. We propose an original computer-aided diagnosis system returning a malignancy score to any suspicious region outlined on MR images, which can be used as a second view by radiologists. The CAD performances are evaluated based on a clinical database of 30 patients, exhaustively and reliably annotated thanks to the histological ground truth obtained via prostatectomy. Finally, we demonstrate the influence of this system in clinical condition based on a ROC analysis involving 12 radiologists, and show a significant increase of diagnostic accuracy, rating confidence and a decrease in inter-expert variability. Building an anatomo-radiological correlation database is a complex and fastidious task, so that numerous studies base their evaluation analysis on the expertise of one experienced radiologist, which is thus doomed to contain uncertainties. We propose a new classification scheme, based on the support vector machine (SVM) algorithm, which is able to account for uncertain data during the learning step. The results obtained, both on toy examples and on our clinical database, demonstrate the potential of this new approach that can be extended to any machine learning problem relying on a probabilitic labelled dataset

Nous présentons deux contributions majeures : 1) une combinaison de plusieurs descripteurs d’images pour la classification à grande échelle, 2) des algorithmes parallèles de SVM pour la classification d’images à grande échelle. Nous proposons aussi un algorithme incrémental et parallèle de classification lorsque les données ne peuvent plus tenir en mémoire vive. / We have proposed a novel method of combination multiple of different features for image classification. For large scale learning classifiers, we have developed the parallel versions of both state-of-the-art linear and nonlinear SVMs. We have also proposed a novel algorithm to extend stochastic gradient descent SVM for large scale learning. A class of large scale incremental SVM classifiers has been developed in order to perform classification tasks on large datasets with very large number of classes and training data can not fit into memory.

Cette thèse s’inscrit dans le contexte de la reconnaissance des activités à partir de séquences vidéo qui est une des préoccupations majeures dans le domaine de la vision par ordinateur. Les domaines d'application pour ces systèmes de vision sont nombreux notamment la vidéo surveillance, la recherche et l'indexation automatique de vidéos ou encore l'assistance aux personnes âgées. Cette tâche reste problématique étant donnée les grandes variations dans la manière de réaliser les activités, l'apparence de la personne et les variations des conditions d'acquisition des activités. L'objectif principal de ce travail de thèse est de proposer une méthode de reconnaissance efficace par rapport aux différents facteurs de variabilité. Les représentations basées sur les points d'intérêt ont montré leur efficacité dans les travaux d'art; elles ont été généralement couplées avec des méthodes de classification globales vue que ses primitives sont temporellement et spatialement désordonnées. Les travaux les plus récents atteignent des performances élevées en modélisant le contexte spatio-temporel des points d'intérêts par exemple certains travaux encodent le voisinage des points d'intérêt à plusieurs échelles. Nous proposons une méthode de reconnaissance des activités qui modélise explicitement l'aspect séquentiel des activités tout en exploitant la robustesse des points d'intérêts dans les conditions réelles. Nous commençons par l'extractivité des points d'intérêt dont a montré leur robustesse par rapport à l'identité de la personne par une étude tensorielle. Ces primitives sont ensuite représentées en tant qu'une séquence de sac de mots (BOW) locaux: la séquence vidéo est segmentée temporellement en utilisant la technique de fenêtre glissante et chacun des segments ainsi obtenu est représenté par BOW des points d'intérêt lui appartenant. Le premier niveau de notre système de classification séquentiel hybride consiste à appliquer les séparateurs à vaste marge (SVM) en tant que classifieur de bas niveau afin de convertir les BOWs locaux en des vecteurs de probabilités des classes d'activité. Les séquences de vecteurs de probabilité ainsi obtenues sot utilisées comme l'entrées de classifieur séquentiel conditionnel champ aléatoire caché (HCRF). Ce dernier permet de classifier d'une manière discriminante les séries temporelles tout en modélisant leurs structures internes via les états cachés. Nous avons évalué notre approche sur des bases publiques ayant des caractéristiques diverses. Les résultats atteints semblent être intéressant par rapport à celles des travaux de l'état de l'art. De plus, nous avons montré que l'utilisation de classifieur de bas niveau permet d'améliorer la performance de système de reconnaissance vue que le classifieur séquentiel HCRF traite directement des informations sémantiques des BOWs locaux, à savoir la probabilité de chacune des activités relativement au segment en question. De plus, les vecteurs de probabilités ont une dimension faible ce qui contribue à éviter le problème de sur apprentissage qui peut intervenir si la dimension de vecteur de caractéristique est plus importante que le nombre des données; ce qui le cas lorsqu'on utilise les BOWs qui sont généralement de dimension élevée. L'estimation les paramètres du HCRF dans un espace de dimension réduite permet aussi de réduire le temps d'entrainement / Human activity recognition (HAR) from video sequences is one of the major active research areas of computer vision. There are numerous application HAR systems, including video-surveillance, search and automatic indexing of videos, and the assistance of frail elderly. This task remains a challenge because of the huge variations in the way of performing activities, in the appearance of the person and in the variation of the acquisition conditions. The main objective of this thesis is to develop an efficient HAR method that is robust to different sources of variability. Approaches based on interest points have shown excellent state-of-the-art performance over the past years. They are generally related to global classification methods as these primitives are temporally and spatially disordered. More recent studies have achieved a high performance by modeling the spatial and temporal context of interest points by encoding, for instance, the neighborhood of the interest points over several scales. In this thesis, we propose a method of activity recognition based on a hybrid model Support Vector Machine - Hidden Conditional Random Field (SVM-HCRF) that models the sequential aspect of activities while exploiting the robustness of interest points in real conditions. We first extract the interest points and show their robustness with respect to the person's identity by a multilinear tensor analysis. These primitives are then represented as a sequence of local "Bags of Words" (BOW): The video is temporally fragmented using the sliding window technique and each of the segments thus obtained is represented by the BOW of interest points belonging to it. The first layer of our hybrid sequential classification system is a Support Vector Machine that converts each local BOW extracted from the video sequence into a vector of activity classes’ probabilities. The sequence of probability vectors thus obtained is used as input of the HCRF. The latter permits a discriminative classification of time series while modeling their internal structures via the hidden states. We have evaluated our approach on various human activity datasets. The results achieved are competitive with those of the current state of art. We have demonstrated, in fact, that the use of a low-level classifier (SVM) improves the performance of the recognition system since the sequential classifier HCRF directly exploits the semantic information from local BOWs, namely the probability of each activity relatively to the current local segment, rather than mere raw information from interest points. Furthermore, the probability vectors have a low-dimension which prevents significantly the risk of overfitting that can occur if the feature vector dimension is relatively high with respect to the training data size; this is precisely the case when using BOWs that generally have a very high dimension. The estimation of the HCRF parameters in a low dimension allows also to significantly reduce the duration of the HCRF training phase

Les Interfaces Cerveau-Machine (ICM) nécessitent l'utilisation de méthodes d'apprentissage statistique pour la reconnaissance de signaux. Dans cette thèse, nous proposons une approche générale permettant d'intégrer des connaissances a priori dans le processus d'apprentissage. Cette approche consiste à apprendre de manière jointe le classifieur et la représentation des données lors d'une optimisation unique. Nous nous sommes plus particulièrement intéressés à des problèmes de sélection de capteurs et proposons plusieurs termes de régularisation adaptés pour ces problèmes. Notre première contribution est une méthode d'apprentissage supervisé de filtres: le filtrage vaste marge. Un filtrage maximisant la marge entre les échantillons est appris et permet de s'adapter automatiquement aux caractéristiques des signaux tout en restant interprétable. Une application ICM et une extension 2D du filtrage a été réalisée. La seconde contribution est une méthode d'apprentissage multitâche parcimonieuse. Elle permet de sélectionner de manière jointe un ensemble de noyaux pertinents pour l'ensemble des tâches de classification. Des algorithmes efficaces ont été proposés pour résoudre le problème d'optimisation et des expérimentations numériques ont montré l'intérêt de l'approche. Finalement, la troisième contribution est une application de l'apprentissage multitâche parcimonieux sur un ensemble de jeux de données ICM. Un terme de régularisation plus général permettant de promouvoir une similarité entre classifieurs est également proposé. Les résultats numériques ont montré qu'une réduction importante du temps de calibration peut être obtenue grâce à l'apprentissage multitâche proposé.

According to the Sectional Titles Act 95 of 1986, levies are assigned according to the size of a unit, in other words a unit’s participation quota. These levies are used to finance a complex’s insurance, common property electricity and water, lift maintenance, audit fees, management agent fees, salaries and wages, security, swimming pool expenses and general building maintenance. According to the Traditional Costing System, indirect costs are allocated based on a single cost actual expenses over a year are dissected. A regression and correlation analysis was done on the relationship between costs, participation quota and levies. The contribution of the study is that it empirically determines the behaviour of cost items in order to allocate indirect costs more accurately. Consequently, this will result in improved reasonability and will eliminate the cross-subsidisation of units. The objectives were reached as follows: The first objective was reached when it was proved through empirical studies that the Sectional Title is not fair when it states that the allocation of levies must be according to floor space. The second objective was reached when both participation quota and units per complex were identified as cost drivers for each cost item. The third objective was reached by allocating the levy on a 50 percent base according to units and a 50 percent base according to the total area. This could be refined by each complex according to the unique need of the complex. driver. However, various indirect costs are not necessarily subject to this single cost driver. In practice, it has been proven that high volume products are proportionally taxed with indirect costs when compared to low volume products. In many complexes, owners with larger units are of the opinion that their larger units are subsidising those owners with smaller units. In contrast to the traditional costing system, activity-based costing recognises that indirect costs can also be assigned by use of multiple cost drivers. Since these multiple cost drivers – which affect the way costs are assigned – can be identified, indirect costs are assigned more accurately. This study investigates the current method of cost allocation, whereby costs are assigned via participation quota, and makes suggestions on how these costs can be assigned on a more accurate and fair basis in practice. The goal of this study is to challenge the reasonability of the Sectional Titles Act. The objectives of the dissertation are firstly, to determine the cost behaviour of various cost items and whether the participation quota is a fair cost driver for determining the levies that are to be paid; secondly, it investigates alternative cost drivers that will be more applicable to certain cost items; thirdly, to suggest a cost formula to replace the current cost method, that being the allocation of indirect costs via participation quota. Empirical methods have been used in the research. The empirical research was performed using data obtained from ANGOR Property Specialists (Pty) Ltd’s database. A sample of 113 complexes was extracted from the database of which the / Thesis (M.Com. (Management Accountancy))--North-West University, Potchefstroom Campus, 2011