自民國69年政府開放國人出國觀光之後，由於國民所得的提高、台幣的升 值及其它種種社經有利因素的影響，使得每年出國觀光人數穩定的成長， 而在民國76年開放國人赴大陸探親之後，出國觀光人數更呈直線上升，這 對於提高國家知名度以及展示國家整體經濟實力有極為明顯的助益。出國 觀光旅客人數的多寡直接或間接影響本地觀光業者及政府相關單位對觀光 業軟、硬體設施的投資以及整體策略的規劃，舉凡國際航線的開拓、航空 公司航線的增減、導遊人員的培訓以及政府駐外單位的配合措施，在在都 有賴於對未來需求的精確預測，過於粗略或不當的預測，不僅將造成大量 觀光資源的閒置與浪費，也將使得政府與觀光業者在這場日趨激烈的觀光 事業競爭中處於極不利的地位。本研究搜集並參考近十年來國內外學者在 觀光旅遊預測模式方面的研究，針對出國觀光旅客整體及各主要市場需求 ，尋找並建立適當之長短期預測模式。我們考慮下列六種模式：簡算法、 單變量時間序列模式、轉移函數模式、時間趨勢模式、指數平滑法以及計 量經濟模式，同時利用各類模式選取準則如AIC、SBC等來選取最佳模式， 或以平均絕對百分誤差(MA PE)、根均方百分誤差(RMSPE)、方向變化誤 差(Direction of Change Erro r)以及趨勢變化誤差(Trend Change Error)來評估各模式預測能力，從中選出最佳模式並進行預測整合分析。

El foc bacterià és una malaltia que afecta a plantes de la família de la rosàcies, causada pel bacteri Erwinia amylovora. El seu rang d'hostes inclou arbres fruiters, com la perera, la pomera o el codonyer, i plantes ornamentals de gran interès comercial i econòmic. Actualment, la malaltia s'ha dispersat i es troba àmpliament distribuïda en totes les zones de clima temperat del món. A Espanya, on la malaltia no és endèmica, el foc bacterià es va detectar per primer cop al 1995 al nord del país (Euskadi) i posteriorment, han aparegut varis focus en altres localitzacions, que han estat convenientment eradicats. El control del foc bacterià, és molt poc efectiu en plantes afectades per la malaltia, de manera que es basa en mesures encaminades a evitar la dispersió del patogen, i la introducció de la malaltia en regions no endèmiques. En aquest treball, la termoteràpia ha estat avaluada com a mètode d'eradicació d'E. amylovora de material vegetal de propagació asimptomàtic. S'ha demostrat que la termoteràpia és un mètode viable d'eradicar E. amylovora de material de propagació. Gairebé totes les espècies i varietats de rosàcies mantingudes en condicions d'humitat sobrevivien 7 hores a 45 ºC i més de 3 hores a 50 ºC, mentre que més d'1 hora d'exposició a 50 ºC amb calor seca produïa danys en el material vegetal i reduïa la brotació. Tractaments de 60 min a 45 ºC o 30 min a 50 ºC van ser suficients per reduir la població epífita d'E. amylovora a nivells no detectables (5 x 102 ufc g-1 p.f.) en branques de perera. Els derivats dels fosfonats i el benzotiadiazol són efectius en el control del foc bacterià en perera i pomera, tant en condicions de laboratori, com d'hivernacle i camp. Els inductors de defensa de les plantes redueixen els nivells de malaltia fins al 40-60%. Els intervals de temps mínims per aconseguir el millor control de la malaltia van ser 5 dies pel fosetil-Al, i 7 dies per l'etefon i el benzotiadiazol, i les dosis òptimes pel fosetil-Al i el benzotiadiazol van ser 3.72 g HPO32- L-1 i 150 mg i.a. L-1, respectivament. Es millora l'eficàcia del fosetil-Al i del benzotiadiazol en el control del foc bacterià, quan es combinen amb els antibiòtics a la meitat de la dosi d'aquests últims. Tot i que l'estratègia de barrejar productes és més pràctica i fàcil de dur a terme a camp, que l'estratègia de combinar productes, el millor nivell de control de la malaltia s'aconsegueix amb l'estratègia de combinar productes. Es va analitzar a nivell histològic i ultrastructural l'efecte del benzotiadiazol i dels fosfonats en la interacció Erwinia amylovora-perera. Ni el benzotiadiazol, ni el fosetil-Al, ni l'etefon van induir canvis estructurals en els teixits de perera 7 dies després de la seva aplicació. No obstant, després de la inoculació d'E. amylovora es va observar en plantes tractades amb fosetil-Al i etefon una desorganització estructural cel·lular, mentre que en les plantes tractades amb benzotiadiazol aquestes alteracions tissulars van ser retardades. S'han avaluat dos models (Maryblyt, Cougarblight) en un camp a Espanya afectat per la malaltia, per determinar la precisió de les prediccions. Es van utilitzar dos models per elaborar el mapa de risc, el BRS-Powell combinat i el BIS95 modificat. Els resultats van mostrar dos zones amb elevat i baix risc de la malaltia. Maryblyt i Cougarblight són dos models de fàcil ús, tot i que la seva implementació en programes de maneig de la malaltia requereix que siguin avaluats i validats per un període de temps més llarg i en àrees on la malaltia hi estigui present. / Fire blight, caused by the bacterium Erwinia amylovora, is a serious disease of rosaceous plants that affects fruit trees such as pear, apple or quince, and ornamental plants with great commercial and economic interest. The disease is spread and well distributed in all temperate regions of the world. In Spain, where the disease is non endemic, fire blight was first detected in 1995 in the North of the country (Euskadi) and later, several new outbreaks have appeared in other locations that have been properly eradicated. Control of fire blight is very slightly effective in affected plants and is based on measures to avoid the spread of pathogen, and the introduction of disease in non-endemic regions. In this work, thermotherapy has been evaluated as a method for eradication of E. amylovora from symptomless propagating plant material. It has been demonstrated that heat is a viable method for eradicating E. amylovora from the propagation material of the pear. Almost all rosaceous species and cultivars maintained under moist conditions survived 7 hours at 45 ºC and up to 3 hours at 50 ºC, while more than 1 hour of exposure at 50 ºC under dry heat injured plants and reduced shooting. However, 60 min at 45 ºC or 30 min at 50 ºC were enough to reduce epiphytic E. amylovora population on pear budwoods to non-detectable level (5 x 102 cfu g-1 f.w.). Phosphonate derivatives and benzothiadiazole were effective in fire blight control in pear and apple, under laboratory, greenhouse and field conditions. Plant defense inducers reduced disease levels to 40-60%. The minimal time intervals to achieve the best control of disease were 5 days for fosetyl-Al, and 7 days for ethephon and benzothiadiazole, and the optimal doses of fosetyl-Al and benzothiadiazole were 3.72 g HPO32- L-1 and 150 mg a.i. L-1, respectively. The efficacy of fosetyl-Al and benzothiadiazole in fire blight control was improved when consecutively sprayed (combined strategy) with a half-reduced dose of antibiotics. Although the mixed strategy is more practical and easier to apply in the orchard than the combined one, the best level of fire blight control was achieved with the combined strategy. The effect of benzothiadiazole and phosphonates in Erwinia amylovora-pear interaction was analyzed at histological and ultrastructural level. Neither benzothiadiazole, nor fosetyl-Al, nor ethephon induced structural changes in pear leaf tissues 7 days after their application. However, after E. amylovora inoculation structural cell disorganization was observed in fosetyl-Al and ethephon-sprayed plants, while in benzothiadiazole-sprayed plants these tissue alterations were delayed. Two predictive models (Maryblyt and Cougarblight) were evaluated in an orchard naturally affected by fire blight in Spain, to determine the accuracy of the predictions. The combined BRS-Powell model and the modified BIS95 model were also evaluated. Results showed two clearly differentiated geographical areas with high and low fire blight risk. Maryblyt and Cougarblight are easy models to use, but their implementation in disease management programs must be evaluated and validated for more seasons and in areas where the disease is present.

No description available.

La cécidomyie orange du blé, Sitodiplosis mosellana (Géhin), est un ravageur commun du froment. Présente sur les trois continents de l’hémisphère Nord, cette espèce est connue depuis deux siècles mais son contrôle reste difficile tant par sa présence discrète que par ses effectifs hautement variables. En Wallonie, les niveaux d’infestations sont globalement faibles mais atteignent localement des seuils inquiétants.<p>Au début des années 2000, la phéromone sexuelle de S. mosellana a été identifiée au Canada. Cette découverte a permis la fabrication de pièges qui ont considérablement amélioré la détection et la mesure des vols de cet insecte minuscule. De 2007 à 2010, les captures de S. mosellana ont été mesurées quotidiennement dans plusieurs dizaines de champs de Wallonie, aux historiques et aux couverts variés.<p>Les volumes de captures au piège à phéromone sexuelle ont été très importants. Il a fallu en étudier la signification, notamment en termes de mesure du risque. En effet, si les mâles sont efficacement capturés, seules les femelles constituent un risque de dégâts. L’interprétation correcte des captures à l’aide de ce type de piège, a été rendue possible par l’observation de différences fondamentales concernant la mobilité et la distribution spatiale des mâles et des femelles de S. mosellana. Même s’ils ne mesurent pas directement l’émergence proprement dite, les pièges à phéromone ont permis, grâce à leur très grande sensibilité, de préciser les connaissances sur l’émergence des adultes et de révéler que plusieurs vagues d’émergence pouvaient se succéder au cours d’une même année.<p>La prévision des émergences de la cécidomyie orange du blé, constitue la clé de voûte de la lutte contre ce ravageur dont un contrôle efficace par des insecticides ne se justifie éventuellement que lorsque la courte saison des pontes coïncide avec l’épiaison des froments. Les patrons d’émergence obtenus par les pièges ont été confrontés aux prévisions de différents modèles conçus en Europe ou en Amérique du Nord, et appliqués aux conditions météorologiques observées de 2007 à 2010. Aucun de ces modèles n’a prévu correctement les émergences sur l’ensemble des quatre années.<p>Les données d’émergence obtenues à l’aide des pièges à phéromone (effectifs élevés et relevés quotidiens) ont fait apparaître une relation de cause à effet entre, d’une part les vagues d’émergences et, d’autre part les épisodes pluvieux observés trois à six semaines plus tôt. L’écart entre une &61618;pluie inductrice&61618; et la vague d’émergence induite correspondante s’est avéré constant en termes d’accumulation de température :il équivaut à 160 degrés-jours en base 7°C. Partant de ce constat et des acquis des modèles antérieurs, un modèle prévisionnel original des émergences a été développé et validé sur le terrain. Allié à une meilleure connaissance de la biologie du ravageur, il constitue un outil majeur de la lutte intégrée.<p>Par ailleurs, les travaux menés ont également révélé l’existence d’un biais fréquent dans les essais d’évaluation des variétés, entraîné par la concentration des pontes de cécidomyie orange sur les premières parcelles atteignant le stade épiaison. Dans le système d’évaluation en vigueur, notamment pour l’inscription dans les catalogues nationaux, ce biais conduit à une sous-estimation du potentiel de rendement des variétés de blé les plus précoces.<p>Enfin, la découverte du rôle inducteur des pluies sur l’émergence des adultes a été exploitée en conditions contrôlées pour planifier des émergences échelonnées, et pour disposer, pendant une longue période, de jeunes adultes prêts à pondre. Cette application permet dès à présent de mesurer en serre le niveau de résistance des variétés exposées de façon homogène à l’insecte, quel que soit leur degré de précocité.<p>De diverses façons, cette étude contribue à une meilleure connaissance de la cécidomyie orange du blé et offre de nouveaux outils pour la lutte intégrée contre ce ravageur.<p><p>--------------------------------------------------<p><p>The orange wheat blossom midge, Sitodiplosis mosellana (Géhin), a common pest of wheat throughout the northern hemisphere, is known for two centuries but remains difficult to control due to its discrete behavior and its highly variable population level.<p>In general, the infestation levels in Wallonia (Belgium) are low, although levels could locally exceed worrying thresholds. <p>In the early 2000s, the sexual pheromone of S. mosellana has been identified in Canada. This discovery has led to the manufacturing of traps which have greatly improved the detection of this tiny insect. From 2007 to 2010 in Wallonia, S. mosellana captures have been daily registered in about 20 fields with different cropping histories and grown with different crops.<p>Insect captures by pheromone traps were numerous. Relation between amount of captures and risk measurement has been studied. As expected, only the males are attracted by the pheromone and the risk of ears infestation is mainly related to the presence of females. The correct interpretation of captures in pheromone traps has been established by the observation of fundamental differences between males and females concerning their mobility and their spatial distribution. <p>Despite the fact that pheromone traps are not real emergence traps, they have led to new information on adult emergence indicating that several emergence waves can be consecutive during the same year. <p>Forecasting the emergence of the adult orange wheat blossom midge is a key element on pest management. Insecticides treatments are sometimes justified when the egg laying period of the insect coincides with ear emergence of wheat. Emergence patterns established from captures of pheromone traps have been compared with the forecast of several models built in Europe or North America. These forecasting models were used with the meteorological data observed from 2007 until 2010. None of the six tested models provided a reliable forecast across the four years of our study. <p>Emergence data from catches in pheromone traps were very accurate because the number of catches were high and were taken each day. This emergence data showed a relation between emergence waves and rainfalls occurring during the preceding 3 to 6 weeks. The lag between inductive rain and emergence wave is constant in terms of temperature accumulation: it is equivalent to 160 degree–days above 7°C. This discovery, combined with experience from previous models, was incorporated into a new forecasting model.<p>In addition, the present work has also revealed the existence of a common bias in variety evaluation trials leading sometime to the concentration of the eggs in the earliest earing variety. In the current evaluation system, this bias leads to a sub-evaluation of the yield for the most precocious varieties. <p>Finally, the discovery of the inductive rain for adult emergence has been used in the screening for resistant varieties to S. mosellana by providing adults during the complete duration of the test. This application of the model allows to measure, under controlled conditions, the level of resistance of all varieties (early and late heading varieties) which are exposed homogeneously to the insect.<p>In total, this study has contributed to a better understanding of the orange wheat blossom midge and provides some new tools in the management of this pest.<p> / Doctorat en Sciences agronomiques et ingénierie biologique / info:eu-repo/semantics/nonPublished

Time series forecasting is a fundamental prerequisite for decision-making processes and crucial in a number of domains such as production planning and energy load balancing. In the past, forecasting was often performed by statistical experts in dedicated software environments outside of current database systems. However, forecasts are increasingly required by non-expert users or have to be computed fully automatically without any human intervention. Furthermore, we can observe an ever increasing data volume and the need for accurate and timely forecasts over large multi-dimensional data sets. As most data subject to analysis is stored in database management systems, a rising trend addresses the integration of forecasting inside a DBMS. Yet, many existing approaches follow a black-box style and try to keep changes to the database system as minimal as possible. While such approaches are more general and easier to realize, they miss significant opportunities for improved performance and usability. In this thesis, we introduce a novel approach that seamlessly integrates time series forecasting into a traditional database management system. In contrast to flash-back queries that allow a view on the data in the past, we have developed a Flash-Forward Database System (F2DB) that provides a view on the data in the future. It supports a new query type - a forecast query - that enables forecasting of time series data and is automatically and transparently processed by the core engine of an existing DBMS. We discuss necessary extensions to the parser, optimizer, and executor of a traditional DBMS. We furthermore introduce various optimization techniques for three different types of forecast queries: ad-hoc queries, recurring queries, and continuous queries. First, we ease the expensive model creation step of ad-hoc forecast queries by reducing the amount of processed data with traditional sampling techniques. Second, we decrease the runtime of recurring forecast queries by materializing models in a specialized index structure. However, a large number of time series as well as high model creation and maintenance costs require a careful selection of such models. Therefore, we propose a model configuration advisor that determines a set of forecast models for a given query workload and multi-dimensional data set. Finally, we extend forecast queries with continuous aspects allowing an application to register a query once at our system. As new time series values arrive, we send notifications to the application based on predefined time and accuracy constraints. All of our optimization approaches intend to increase the efficiency of forecast queries while ensuring high forecast accuracy.

The ability of generating precise numerical forecasts is important to successful Enterprises in order to prepare themselves for undetermined future developments. For Utility companies, forecasts of prospective energy demand are a crucial component in order to maintain the physical stability and reliability of electricity grids. The constantly increasing capacity of fluctuating renewable energy sources creates a challenge in balancing power supply and demand. To allow for better integration, supply forecasting has become an important topic in the research field of energy data management and many new forecasting methods have been proposed in the literature. However, choosing the optimal solution for a specific forecasting problem remains a time- and work-intensive Task as meaningful benchmarks are rare and there is still no standard, easy-to-use, and robust approach. Many of the models in use are obtained by executing black-box machine learning tools and then manually optimized by human experts via trial-and-error towards the requirements of the underlying use case. Due to the lack of standardized Evaluation methodologies and access to experimental data, these results are not easily comparable. In this thesis, we address the topic of systematic benchmarks for renewable Energy supply forecasts. These usually include two stages, requiring a weather- and an energy forecast model. The latter can be selected amongst the classes of physical, statistical, and hybrid models. The selection of an appropriate model is one of the major tasks included in the forecasting process. We conducted an empirical analysis to assess the most popular forecasting methods. In contrast to the classical time- and resource intensive, mostly manual evaluation procedure, we developed a more efficient decision-support solution. With the inclusion of contextual information, our heuristic approach HMR is able to identify suitable examples in a case base and generates a recommendation out of the results from already existing solutions. The usage of time series representations reduces the dimensions of the original data thus allowing for an efficient search in large data sets. A context-aware evaluation methodology is introduced to assess a forecast’s quality based on its monetary return in the corresponding market environment. Results otherwise usually evaluated using statistical accuracy criteria become more interpretable by estimating real-world impacts. Finally, we introduced the ECAST framework as an open and easy to-use online platform that supports the benchmarking of energy time series forecasting methods. It aides inexperienced practitioners by supporting the execution of automated tasks, thus making complex benchmarks much more efficient and easy to handle. The integration of modules like the Ensembler, the Recommender, and the Evaluator provide additional value for forecasters. Reliable benchmarks can be conducted on this basis, while analytical functions for output explanation provide transparency for the user.:1 INTRODUCTION 11 2 ENERGY DATA MANAGEMENT CHALLENGES 17 2.1 Market Relevance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.2 EDMS Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.2.1 Core Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.2.2 Typical Energy Data Management Processes . . . . . . . . . . . 23 2.2.3 System Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.3 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.3.1 Smart Metering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.3.2 Energy Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.3.3 Energy Saving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.3.4 Mobile Consumption Devices . . . . . . . . . . . . . . . . . . . . . 30 2.3.5 Smart Grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 3 ENERGY SUPPLY FORECASTING CONCEPTS 35 3.1 Energy Supply Forecasting Approaches . . . . . . . . . . . . . . . . . . . 36 3.1.1 Weather Forecast Models . . . . . . . . . . . . . . . . . . . . . . . . 36 3.1.2 Energy Forecast Models . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.2 Energy Forecasting Process . . . . . . . . . . . . . . . . . . . . . . . . . . 43 3.2.1 Iterative Standard Process Model . . . . . . . . . . . . . . . . . . . 43 3.2.2 Context-Awareness . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.3 Model Selection - A Benchmark Case Study . . . . . . . . . . . . . . . . 48 3.3.1 Use Case Description . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.3.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 3.3.3 Result Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4 RELEVANCE OF RENEWABLE ENERGY FORECASTING METHODS 55 4.1 Scientific Relevance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.1.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.1.2 Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 4.1.3 Qualitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.2 Practical Relevance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.2.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.2.2 Feedback from Software Providers . . . . . . . . . . . . . . . . . . 61 4.2.3 Feedback from Software Users . . . . . . . . . . . . . . . . . . . . . 62 4.3 Forecasting Competitions . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 5 HEURISTIC MODEL RECOMMENDATION 67 5.1 Property-based Similarity Determination . . . . . . . . . . . . . . . . . . 67 5.1.1 Time Series Similarity . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 5.1.2 Reducing Dimensionality with Property Extraction . . . . . . . . . 69 5.1.3 Correlation Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 5.2 Feature Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 5.2.1 Feature Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 5.2.2 Feature Pre-Selection . . . . . . . . . . . . . . . . . . . . . . . . . . 83 5.2.3 Property-based Least Angle Regression . . . . . . . . . . . . . . . 85 5.3 HMR Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 5.3.1 Formalized Foundations . . . . . . . . . . . . . . . . . . . . . . . . . 87 5.3.2 Procedure Description . . . . . . . . . . . . . . . . . . . . . . . . . . 88 5.3.3 Quality Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 5.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 5.4.1 Case Base Composition . . . . . . . . . . . . . . . . . . . . . . . . . 91 5.4.2 Classifier Performance on univariate Models . . . . . . . . . . . . 95 5.4.3 HMR performance on multivariate models . . . . . . . . . . . . . 99 5.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 6 VALUE-BASED RESULT EVALUATION METHODOLOGY 105 6.1 Accuracy evaluation in energy forecasting . . . . . . . . . . . . . . . . 106 6.2 Energy market models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 6.3 Value-based forecasting performance . . . . . . . . . . . . . . . . . . . 110 6.3.1 Forecast Benefit Determination . . . . . . . . . . . . . . . . . . . . 110 6.3.2 Multi-dimensional Ranking Scores . . . . . . . . . . . . . . . . . . . 113 6.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 6.4.1 Use Case Description . . . . . . . . . . . . . . . . . . . . . . . . . . 117 6.4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 6.4.3 Result Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 6.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 7 ECAST BENCHMARK FRAMEWORK 129 7.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 7.1.1 Objective Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 7.1.2 Fundamental Design Principles . . . . . . . . . . . . . . . . . . . . 131 7.2 System Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 7.2.1 Task Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 7.2.2 System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 7.3 Demonstration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 7.3.1 Step 1: Create a new Benchmark . . . . . . . . . . . . . . . . . . 137 7.3.2 Step 2: Build Ensembles . . . . . . . . . . . . . . . . . . . . . . . . . 139 7.3.3 Step 3: Evaluate the Output . . . . . . . . . . . . . . . . . . . . . . 141 7.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 8 CONCLUSIONS 145 BIBLIOGRAPHY 149 LIST OF FIGURES 167 LIST OF TABLES 169 A LIST OF REVIEWED JOURNAL ARTICLES 171 B QUESTIONNAIRES 175 C STANDARD ERRORS FOR RANKING SCORES 179 D ERROR DISTRIBUTION FOR BENCHMARKED PREDICTORS 183