Contextual text mining

With the dramatic growth of text information, there is an increasing need for powerful text mining systems that can automatically discover useful knowledge from text. Text is generally associated with all kinds of contextual information. Those contexts can be explicit, such as the time and the location where a blog article is written, and the author(s) of a biomedical publication, or implicit, such as the positive or negative sentiment that an author had when she wrote a product review; there may also be complex context such as the social network of the authors. Many applications require analysis of topic patterns over different contexts. For instance, analysis of search logs in the context of the user can reveal how we can improve the quality of a search engine by optimizing the search results according to particular users; analysis of customer reviews in the context of positive and negative sentiments can help the user summarize public opinions about a product; analysis of blogs or scientific publications in the context of a social network can facilitate discovery of more meaningful topical communities. Since context information significantly affects the choices of topics and language made by authors, in general, it is very important to incorporate it into analyzing and mining text data. In general, modeling the context in text, discovering contextual patterns of language units and topics from text, a general task which we refer to as Contextual Text Mining, has widespread applications in text mining. In this thesis, we provide a novel and systematic study of contextual text mining, which is a new paradigm of text mining treating context information as the ``first-class citizen.'' We formally define the problem of contextual text mining and its basic tasks, and propose a general framework for contextual text mining based on generative modeling of text. This conceptual framework provides general guidance on text mining problems with context information and can be instantiated into many real tasks, including the general problem of contextual topic analysis. We formally present a functional framework for contextual topic analysis, with a general contextual topic model and its various versions, which can effectively solve the text mining problems in a lot of real world applications. We further introduce general components of contextual topic analysis, by adding priors to contextual topic models to incorporate prior knowledge, regularizing contextual topic models with dependency structure of context, and postprocessing contextual patterns to extract refined patterns. The refinements on the general contextual topic model naturally lead to a variety of probabilistic models which incorporate different types of context and various assumptions and constraints. These special versions of the contextual topic model are proved effective in a variety of real applications involving topics and explicit contexts, implicit contexts, and complex contexts. We then introduce a postprocessing procedure for contextual patterns, by generating meaningful labels for multinomial context models. This method provides a general way to interpret text mining results for real users. By applying contextual text mining in the ``context'' of other text information management tasks, including ad hoc text retrieval and web search, we further prove the effectiveness of contextual text mining techniques in a quantitative way with large scale datasets. The framework of contextual text mining not only unifies many explorations of text analysis with context information, but also opens up many new possibilities for future research directions in text mining.

Leveraging Multiple Features for Image Retrieval and Matching

The goal of image retrieval and matching is to find and locate object instances in images from a large-scale image database. While visual features are abundant, how to combine them to improve performance by individual features remains a challenging task. In this work, we focus on leveraging multiple features for accurate and efficient image retrieval and matching. We first propose two graph-based approaches to rerank initially retrieved images for generic image retrieval. In the graph, vertices are images while edges are similarities between image pairs. Our first approach employs a mixture Markov model based on a random walk model on multiple graphs to fuse graphs. We introduce a probabilistic model to compute the importance of each feature for graph fusion under a naive Bayesian formulation, which requires statistics of similarities from a manually labeled dataset containing irrelevant images. To reduce human labeling, we further propose a fully unsupervised reranking algorithm based on a submodular objective function that can be efficiently optimized by greedy algorithm. By maximizing an information gain term over the graph, our submodular function favors a subset of database images that are similar to query images and resemble each other. The function also exploits the rank relationships of images from multiple ranked lists obtained by different features. We then study a more well-defined application, person re-identification, where the database contains labeled images of human bodies captured by multiple cameras. Re-identifications from multiple cameras are regarded as related tasks to exploit shared information. We apply a novel multi-task learning algorithm using both low level features and attributes. A low rank attribute embedding is joint learned within the multi-task learning formulation to embed original binary attributes to a continuous attribute space, where incorrect and incomplete attributes are rectified and recovered. To locate objects in images, we design an object detector based on object proposals and deep convolutional neural networks (CNN) in view of the emergence of deep networks. We improve a Fast RCNN framework and investigate two new strategies to detect objects accurately and efficiently: scale-dependent pooling (SDP) and cascaded rejection classifiers (CRC). The SDP improves detection accuracy by exploiting appropriate convolutional features depending on the scale of input object proposals. The CRC effectively utilizes convolutional features and greatly eliminates negative proposals in a cascaded manner, while maintaining a high recall for true objects. The two strategies together improve the detection accuracy and reduce the computational cost.

Hybrid Variable Neighborhood HyperHeuristics for Exam Timetabling Problems

This paper presents our work on analysing the high level search within a graph based hyperheuristic. The graph based hyperheuristic solves the problem at a higher level by searching through permutations of graph heuristics rather than the actual solutions. The heuristic permutations are then used to construct the solutions. Variable Neighborhood Search, Steepest Descent, Iterated Local Search and Tabu Search are compared. An analysis of their performance within the high level search space of heuristics is also carried out. Experimental results on benchmark exam timetabling problems demonstrate the simplicity and efficiency of this hyperheuristic approach. They also indicate that the choice of the high level search methodology is not crucial and the high level search should explore the heuristic search space as widely as possible within a limited searching time. This simple and general graph based hyperheuristic may be applied to a range of timetabling and optimisation problems.

Hybrid Variable Neighborhood HyperHeuristics for Exam Timetabling Problems

This paper presents our work on analysing the high level search within a graph based hyperheuristic. The graph based hyperheuristic solves the problem at a higher level by searching through permutations of graph heuristics rather than the actual solutions. The heuristic permutations are then used to construct the solutions. Variable Neighborhood Search, Steepest Descent, Iterated Local Search and Tabu Search are compared. An analysis of their performance within the high level search space of heuristics is also carried out. Experimental results on benchmark exam timetabling problems demonstrate the simplicity and efficiency of this hyperheuristic approach. They also indicate that the choice of the high level search methodology is not crucial and the high level search should explore the heuristic search space as widely as possible within a limited searching time. This simple and general graph based hyperheuristic may be applied to a range of timetabling and optimisation problems.

A space-time hybrid hourly rainfall model for derived flood frequency analysis

For derived flood frequency analysis based on hydrological modelling long continuous precipitation time series with high temporal resolution are needed. Often, the observation network with recording rainfall gauges is poor, especially regarding the limited length of the available rainfall time series. Stochastic precipitation synthesis is a good alternative either to extend or to regionalise rainfall series to provide adequate input for long-term rainfall-runoff modelling with subsequent estimation of design floods. Here, a new two step procedure for stochastic synthesis of continuous hourly space-time rainfall is proposed and tested for the extension of short observed precipitation time series. First, a single-site alternating renewal model is presented to simulate independent hourly precipitation time series for several locations. The alternating renewal model describes wet spell durations, dry spell durations and wet spell intensities using univariate frequency distributions separately for two seasons. The dependence between wet spell intensity and duration is accounted for by 2-copulas. For disaggregation of the wet spells into hourly intensities a predefined profile is used. In the second step a multi-site resampling procedure is applied on the synthetic point rainfall event series to reproduce the spatial dependence structure of rainfall. Resampling is carried out successively on all synthetic event series using simulated annealing with an objective function considering three bivariate spatial rainfall characteristics. In a case study synthetic precipitation is generated for some locations with short observation records in two mesoscale catchments of the Bode river basin located in northern Germany. The synthetic rainfall data are then applied for derived flood frequency analysis using the hydrological model HEC-HMS. The results show good performance in reproducing average and extreme rainfall characteristics as well as in reproducing observed flood frequencies. The presented model has the potential to be used for ungauged locations through regionalisation of the model parameters.

The suppliers selection problem: a case study

The effective supplier evaluation and purchasing processes are of vital importance to business organizations, making the suppliers selection problem a fundamental key issue to their success. We consider a complex supplier selection problem with multiple products where minimum package quantities, minimum order values related to delivery costs, and discounted pricing schemes are taken into account. Our main contribution is to present a mixed integer linear programming (MILP) model for this supplier selection problem. The model is used to solve several examples including three real case studies from an electronic equipment assembly company.

The impact of personal background and school contextual factors on academic competence and mental health functioning across the primary-secondary school transition

Students negotiate the transition to secondary school in different ways. While some thrive on the opportunity, others are challenged. A prospective longitudinal design was used to determine the contribution of personal background and school contextual factors on academic competence (AC) and mental health functioning (MHF) of 266 students, 6-months before and after the transition to secondary school. Data from 197 typically developing students and 69 students with a disability were analysed using hierarchical linear regression modelling. Both in primary and secondary school, students with a disability and from socially disadvantaged backgrounds gained poorer scores for AC and MHF than their typically developing and more affluent counterparts. Students who attended independent and mid-range sized primary schools had the highest concurrent AC. Those from independent primary schools had the lowest MHF. The primary school organisational model significantly influenced post-transition AC scores; with students from Kindergarten--Year 7 schools reporting the lowest scores, while those from the Kindergarten--Year 12 structure without middle school having the highest scores. Attending a school which used the Kindergarten--Year 12 with middle school structure was associated with a reduction in AC scores across the transition. Personal background factors accounted for the majority of the variability in post-transition AC and MHF. The contribution of school contextual factors was relatively minor. There is a potential opportunity for schools to provide support to disadvantaged students before the transition to secondary school, as they continue to be at a disadvantage after the transition.

Assessing the impact of agricultural drought on maize prices in Kenya with the approach of the SPOT-VEGETATION NDVI remote sensing

The high cost of maize in Kenya is basically driven by East African regional commodity demand forces and agricultural drought. The production of maize, which is a common staple food in Kenya, is greatly affected by agricultural drought. However, calculations of drought risk and impact on maize production in Kenya is limited by the scarcity of reliable rainfall data. The objective of this study was to apply a novel hyperspectral remote sensing method to modelling temporal fluctuations of maize production and prices in five markets in Kenya. SPOT-VEGETATION NDVI time series were corrected for seasonal effects by computing the standardized NDVI anomalies. The maize residual price time series was further related to the NDVI seasonal anomalies using a multiple linear regression modelling approach. The result shows a moderately strong positive relationship (0.67) between residual price series and global maize prices. Maize prices were high during drought periods (i.e. negative NDVI anomalies) and low during wet seasons (i.e. positive NDVI anomalies). This study concludes that NDVI is a good index for monitoring the evolution of maize prices and food security emergency planning in Kenya. To obtain a very strong correlation for the relationship between the wholesale maize price and the global maize price, future research could consider adding other price-driving factors into the regression models.

Production Diseases in European Organic Dairy Farms – Current Status and Identification of Drivers for Improvement by Means of a Systemic Approach

The aims of this thesis were to determine the animal health status in organic dairy farms in Europe and to identify drivers for improving the current situation by means of a systemic approach. Prevalences of production diseases were determined in 192 herds in Germany, France, Spain, and Sweden (Paper I), and stakeholder consultations were performed to investigate potential drivers to improve animal health on the sector level (ibid.). Interactions between farm variables were assessed through impact analysis and evaluated to identify general system behaviour and classify components according to their outgoing and incoming impacts (Paper II-III). The mean values and variances of prevalences indicate that the common rules of organic dairy farming in Europe do not result in consistently low levels of production diseases. Stakeholders deemed it necessary to improve the current status and were generally in favour of establishing thresholds for the prevalence of production diseases in organic dairy herds as well as taking actions to improve farms below that threshold. In order to close the gap between the organic principle of health and the organic farming practice, there is the need to formulate a common objective of good animal health and to install instruments to ensure and prove that the aim is followed by all dairy farmers in Europe who sell their products under the organic label. Regular monitoring and evaluation of herd health performance based on reference values are considered preconditions for identifying farms not reaching the target and thus in need of improvement. Graph-based impact analysis was shown to be a suitable method for modeling and evaluating the manifold interactions between farm factors and for identifying the most influential components on the farm level taking into account direct and indirect impacts as well as impact strengths. Variables likely to affect the system as a whole, and the prevalence of production diseases in particular, varied largely between farms despite some general tendencies. This finding reflects the diversity of farm systems and underlines the importance of applying systemic approaches in health management. Reducing the complexity of farm systems and indicating farm-specific drivers, i.e. areas in a farm, where changes will have a large impact, the presented approach has the potential to complement and enrich current advisory practice and to support farmers’ decision-making in terms of animal health.

Bidding and Optimization Strategies for Wind-PV Systems in Electricity Markets

The variability in non-dispatchable power generation raises important challenges to the integration of renewable energy sources into the electricity power grid. This paper provides the coordinated trading of wind and photovoltaic energy to mitigate risks due to the wind and solar power variability, electricity prices, and financial penalties arising out the generation shortfall and surplus. The problem of wind-photovoltaic coordinated trading is formulated as a linear programming problem. The goal is to obtain the optimal bidding strategy that maximizes the total profit. The wind-photovoltaic coordinated operation is modeled and compared with the uncoordinated operation. A comparison of the models and relevant conclusions are drawn from an illustrative case study of the Iberian day-ahead electricity market.

Bidding and Optimization Strategies for Wind-PV Systems in Electricity Markets Assisted by CPS

The variability in non-dispatchable power generation raises important challenges to the integration of renewable energy sources into the electricity power grid. This paper provides the coordinated trading of wind and photovoltaic energy assisted by a cyber-physical system for supporting management decisions to mitigate risks due to the wind and solar power variability, electricity prices, and financial penalties arising out the generation shortfall and surplus. The problem of wind-photovoltaic coordinated trading is formulated as a stochastic linear programming problem. The goal is to obtain the optimal bidding strategy that maximizes the total profit. The wind-photovoltaic coordinated operation is modelled and compared with the uncoordinated operation. A comparison of the models and relevant conclusions are drawn from an illustrative case study of the Iberian day-ahead electricity market.

Bidding Decision of Wind-Thermal GenCo in Day-Ahead Market

This paper deals with the self-scheduling problem of a price-taker having wind and thermal power production and assisted by a cyber-physical system for supporting management decisions in a day-ahead electric energy market. The self-scheduling is regarded as a stochastic mixed-integer linear programming problem. Uncertainties on electricity price and wind power are considered through a set of scenarios. Thermal units are modelled by start-up and variable costs, furthermore constraints are considered, such as: ramp up/down and minimum up/down time limits. The stochastic mixed-integer linear programming problem allows a decision support for strategies advantaging from an effective wind and thermal mixed bidding. A case study is presented using data from the Iberian electricity market.

Impactes do Preço da Água na Agricultura no Perímetro Irrigado do Vale de Caxito

O aumento da pressão sobre os recursos hídricos tem levado muitos países a reconsiderarem os mecanismos utilizados na indução do uso eficiente da água, especialmente na agricultura irrigada. Estabelecer o preço correto da água é um dos mecanismos de tornar mais eficiente a alocação da água. O presente trabalho tem como objetivo a análise dos impactes económicos, sociais e ambientais de políticas de preço da água. A metodologia utilizada foi a Programação Linear, aplicada ao Perímetro Irrigado do Vale de Caxito, Província do Bengo, a 45 km de Luanda, que tem como fonte o rio Dande. Foram testados três cenários relativos a políticas de tarifação de água: tarifa volumétrica simples, tarifa volumétrica variável, e tarifa fixa por superfície. As principais conclusões mostram que, do ponto de vista do uso eficiente da água na agricultura, os melhores resultados obtêm-se com a tarifa volumétrica variável; do ponto de vista social, a tarifação volumétrica simples apresenta os melhores resultados; o método de tarifa volumétrica variável foi o mais penalizador, reduzindo rapidamente a área das culturas mais consumidoras de água, sendo o melhor do ponto de vista ambiental. Qualquer um dos métodos traz aspetos negativos relativamente à redução da margem bruta total. Palavras-chaves: Recursos hídricos; Preço da água; Programação linear. Abstract: Increased pressure on water resources has led many countries to reconsider the mechanisms used in the induction of efficient water use, especially for irrigated agriculture, a major consumer of water. Establishing the correct price of water is one of the mechanisms for more efficient allocation of water. This paper aims to analyze the economic, social and essenenvironmental impacts of water price policies. The methodology used is the linear programming, applied to the Irrigated Valley Caxito, in Bengo Province, 45 kilometers from Luanda, which has the river Dande as its source. Three scenarios concerning water price policies were tested: simple volumetric rate, variable volumetric rate and flat rate per surface. The main findings show that from the point of view of the efficient use of water in agriculture, the best results are obtained with variable volumetric rate; from the social point of view, the simple volumetric rate has the best results; the volume variable rate method proved to be the most penalizing, quickly reducing the area of most water consuming cultures, being the method in which the environmental objectives would be more readily achieved. Either methods bring negative aspects in relation to the reduction of total gross margin. Key-words: Water resources; Water price; Linear programming.