Multi-dimensional classification using Bayesian networks for stationary and evolving streaming data

Hoy en día, con la evolución continua y rápida de las tecnologías de la información y los dispositivos de computación, se recogen y almacenan continuamente grandes volúmenes de datos en distintos dominios y a través de diversas aplicaciones del mundo real. La extracción de conocimiento útil de una cantidad tan enorme de datos no se puede realizar habitualmente de forma manual, y requiere el uso de técnicas adecuadas de aprendizaje automático y de minería de datos. La clasificación es una de las técnicas más importantes que ha sido aplicada con éxito a varias áreas. En general, la clasificación se compone de dos pasos principales: en primer lugar, aprender un modelo de clasificación o clasificador a partir de un conjunto de datos de entrenamiento, y en segundo lugar, clasificar las nuevas instancias de datos utilizando el clasificador aprendido. La clasificación es supervisada cuando todas las etiquetas están presentes en los datos de entrenamiento (es decir, datos completamente etiquetados), semi-supervisada cuando sólo algunas etiquetas son conocidas (es decir, datos parcialmente etiquetados), y no supervisada cuando todas las etiquetas están ausentes en los datos de entrenamiento (es decir, datos no etiquetados). Además, aparte de esta taxonomía, el problema de clasificación se puede categorizar en unidimensional o multidimensional en función del número de variables clase, una o más, respectivamente; o también puede ser categorizado en estacionario o cambiante con el tiempo en función de las características de los datos y de la tasa de cambio subyacente. A lo largo de esta tesis, tratamos el problema de clasificación desde tres perspectivas diferentes, a saber, clasificación supervisada multidimensional estacionaria, clasificación semisupervisada unidimensional cambiante con el tiempo, y clasificación supervisada multidimensional cambiante con el tiempo. Para llevar a cabo esta tarea, hemos usado básicamente los clasificadores Bayesianos como modelos. La primera contribución, dirigiéndose al problema de clasificación supervisada multidimensional estacionaria, se compone de dos nuevos métodos de aprendizaje de clasificadores Bayesianos multidimensionales a partir de datos estacionarios. Los métodos se proponen desde dos puntos de vista diferentes. El primer método, denominado CB-MBC, se basa en una estrategia de envoltura de selección de variables que es voraz y hacia delante, mientras que el segundo, denominado MB-MBC, es una estrategia de filtrado de variables con una aproximación basada en restricciones y en el manto de Markov. Ambos métodos han sido aplicados a dos problemas reales importantes, a saber, la predicción de los inhibidores de la transcriptasa inversa y de la proteasa para el problema de infección por el virus de la inmunodeficiencia humana tipo 1 (HIV-1), y la predicción del European Quality of Life-5 Dimensions (EQ-5D) a partir de los cuestionarios de la enfermedad de Parkinson con 39 ítems (PDQ-39). El estudio experimental incluye comparaciones de CB-MBC y MB-MBC con los métodos del estado del arte de la clasificación multidimensional, así como con métodos comúnmente utilizados para resolver el problema de predicción de la enfermedad de Parkinson, a saber, la regresión logística multinomial, mínimos cuadrados ordinarios, y mínimas desviaciones absolutas censuradas. En ambas aplicaciones, los resultados han sido prometedores con respecto a la precisión de la clasificación, así como en relación al análisis de las estructuras gráficas que identifican interacciones conocidas y novedosas entre las variables. La segunda contribución, referida al problema de clasificación semi-supervisada unidimensional cambiante con el tiempo, consiste en un método nuevo (CPL-DS) para clasificar flujos de datos parcialmente etiquetados. Los flujos de datos difieren de los conjuntos de datos estacionarios en su proceso de generación muy rápido y en su aspecto de cambio de concepto. Es decir, los conceptos aprendidos y/o la distribución subyacente están probablemente cambiando y evolucionando en el tiempo, lo que hace que el modelo de clasificación actual sea obsoleto y deba ser actualizado. CPL-DS utiliza la divergencia de Kullback-Leibler y el método de bootstrapping para cuantificar y detectar tres tipos posibles de cambio: en las predictoras, en la a posteriori de la clase o en ambas. Después, si se detecta cualquier cambio, un nuevo modelo de clasificación se aprende usando el algoritmo EM; si no, el modelo de clasificación actual se mantiene sin modificaciones. CPL-DS es general, ya que puede ser aplicado a varios modelos de clasificación. Usando dos modelos diferentes, el clasificador naive Bayes y la regresión logística, CPL-DS se ha probado con flujos de datos sintéticos y también se ha aplicado al problema real de la detección de código malware, en el cual los nuevos ficheros recibidos deben ser continuamente clasificados en malware o goodware. Los resultados experimentales muestran que nuestro método es efectivo para la detección de diferentes tipos de cambio a partir de los flujos de datos parcialmente etiquetados y también tiene una buena precisión de la clasificación. Finalmente, la tercera contribución, sobre el problema de clasificación supervisada multidimensional cambiante con el tiempo, consiste en dos métodos adaptativos, a saber, Locally Adpative-MB-MBC (LA-MB-MBC) y Globally Adpative-MB-MBC (GA-MB-MBC). Ambos métodos monitorizan el cambio de concepto a lo largo del tiempo utilizando la log-verosimilitud media como métrica y el test de Page-Hinkley. Luego, si se detecta un cambio de concepto, LA-MB-MBC adapta el actual clasificador Bayesiano multidimensional localmente alrededor de cada nodo cambiado, mientras que GA-MB-MBC aprende un nuevo clasificador Bayesiano multidimensional. El estudio experimental realizado usando flujos de datos sintéticos multidimensionales indica los méritos de los métodos adaptativos propuestos. ABSTRACT Nowadays, with the ongoing and rapid evolution of information technology and computing devices, large volumes of data are continuously collected and stored in different domains and through various real-world applications. Extracting useful knowledge from such a huge amount of data usually cannot be performed manually, and requires the use of adequate machine learning and data mining techniques. Classification is one of the most important techniques that has been successfully applied to several areas. Roughly speaking, classification consists of two main steps: first, learn a classification model or classifier from an available training data, and secondly, classify the new incoming unseen data instances using the learned classifier. Classification is supervised when the whole class values are present in the training data (i.e., fully labeled data), semi-supervised when only some class values are known (i.e., partially labeled data), and unsupervised when the whole class values are missing in the training data (i.e., unlabeled data). In addition, besides this taxonomy, the classification problem can be categorized into uni-dimensional or multi-dimensional depending on the number of class variables, one or more, respectively; or can be also categorized into stationary or streaming depending on the characteristics of the data and the rate of change underlying it. Through this thesis, we deal with the classification problem under three different settings, namely, supervised multi-dimensional stationary classification, semi-supervised unidimensional streaming classification, and supervised multi-dimensional streaming classification. To accomplish this task, we basically used Bayesian network classifiers as models. The first contribution, addressing the supervised multi-dimensional stationary classification problem, consists of two new methods for learning multi-dimensional Bayesian network classifiers from stationary data. They are proposed from two different points of view. The first method, named CB-MBC, is based on a wrapper greedy forward selection approach, while the second one, named MB-MBC, is a filter constraint-based approach based on Markov blankets. Both methods are applied to two important real-world problems, namely, the prediction of the human immunodeficiency virus type 1 (HIV-1) reverse transcriptase and protease inhibitors, and the prediction of the European Quality of Life-5 Dimensions (EQ-5D) from 39-item Parkinson’s Disease Questionnaire (PDQ-39). The experimental study includes comparisons of CB-MBC and MB-MBC against state-of-the-art multi-dimensional classification methods, as well as against commonly used methods for solving the Parkinson’s disease prediction problem, namely, multinomial logistic regression, ordinary least squares, and censored least absolute deviations. For both considered case studies, results are promising in terms of classification accuracy as well as regarding the analysis of the learned MBC graphical structures identifying known and novel interactions among variables. The second contribution, addressing the semi-supervised uni-dimensional streaming classification problem, consists of a novel method (CPL-DS) for classifying partially labeled data streams. Data streams differ from the stationary data sets by their highly rapid generation process and their concept-drifting aspect. That is, the learned concepts and/or the underlying distribution are likely changing and evolving over time, which makes the current classification model out-of-date requiring to be updated. CPL-DS uses the Kullback-Leibler divergence and bootstrapping method to quantify and detect three possible kinds of drift: feature, conditional or dual. Then, if any occurs, a new classification model is learned using the expectation-maximization algorithm; otherwise, the current classification model is kept unchanged. CPL-DS is general as it can be applied to several classification models. Using two different models, namely, naive Bayes classifier and logistic regression, CPL-DS is tested with synthetic data streams and applied to the real-world problem of malware detection, where the new received files should be continuously classified into malware or goodware. Experimental results show that our approach is effective for detecting different kinds of drift from partially labeled data streams, as well as having a good classification performance. Finally, the third contribution, addressing the supervised multi-dimensional streaming classification problem, consists of two adaptive methods, namely, Locally Adaptive-MB-MBC (LA-MB-MBC) and Globally Adaptive-MB-MBC (GA-MB-MBC). Both methods monitor the concept drift over time using the average log-likelihood score and the Page-Hinkley test. Then, if a drift is detected, LA-MB-MBC adapts the current multi-dimensional Bayesian network classifier locally around each changed node, whereas GA-MB-MBC learns a new multi-dimensional Bayesian network classifier from scratch. Experimental study carried out using synthetic multi-dimensional data streams shows the merits of both proposed adaptive methods.

Sea Bottom Classification by Means of Bathymetric LIDAR Data

Light Detection and Ranging (LIDAR) provides high horizontal and vertical resolution of spatial data located in point cloud images, and is increasingly being used in a number of applications and disciplines, which have concentrated on the exploit and manipulation of the data using mainly its three dimensional nature. Bathymetric LIDAR systems and data are mainly focused to map depths in shallow and clear waters with a high degree of accuracy. Additionally, the backscattering produced by the different materials distributed over the bottom surface causes that the returned intensity signal contains important information about the reflection properties of these materials. Processing conveniently these values using a Simplified Radiative Transfer Model, allows the identification of different sea bottom types. This paper presents an original method for the classification of sea bottom by means of information processing extracted from the images generated through LIDAR data. The results are validated using a vector database containing benthic information derived by marine surveys.

Una aproximación a la integración en Open Data de los recursos INSPIRE de la IDEE.

Este trabajo, «Una aproximación a Ia integración en Open Data de los recursos Inspire de Ia IDEE », tiene por objetivo el construir un puente entre las Infraestructuras de Datos Espaciales (IDE) y el mundo de los «datos abiertos » aprovechando el marco legal de la Reutilización de la Información del Sector Público (RISP). Tras analizar qué es RISP y en particular los datos abiertos, y cómo se implementa en distintas Administraciones, se estudian los requisitos técnicos y legales necesarios para construir el «traductor» que permita canalizar la información IDE en el portal central de reutilización de información español datos.gob.es, dando una mayor visibilidad a los recursos INSPIRE. El trabajo se centra específicamente en dos puntos: en primer lugar en proporcionar y documentar la solución técnica que sirva en primera instancia para que el Instituto Geográfico Nacional aporte con más eficiencia sus recursos a datos.gob.es. En segundo lugar, a estudiar la aplicabilidad de esta misma solución al ámbito de la IDE de España (IDEE), señalando problemas detectados en el análisis de su contenido y sugiriendo recomendaciones para minimizar los problemas de su potencial reutilización. ABSTRACT: This work titled «Analysis of the integration of INSPIRE resources coming from Spanish Spatial Data Infrastructure within the National Public Sector Information portal», aims to build a bridge between the Spatial Data Infrastructures (SDI ) and the world of "Open Data" taking advantage of the legal framework on the Re-use of Public Sector Information (PSI) . After analyzing what PSI reuse and Open Data is and how it is implemented by different administrations, a study to extract the technical and legal requirements is done to build the "translator" that will allow adding SDI resources within the Spanish portal for the PSI reuse data .gob.es while giving greater visibility to INSPIRE. This document specifically focuses on two aspects: first to provide and document the technical solution that serves primarily for the National Geographic Institute to supply more efficiently its resources to datos.gob.es. Secondly, to study the applicability of the proposed solution to the whole Spanish SDI (IDEE), noting identified problems and suggesting recommendations to minimize problems of its potential reuse.

A data model for Cultural Heritage within INSPIRE.

Esta monografía presenta los fundamentos, contexto y detalles técnicos de un Esquema de Aplicación para la incorporación de datos espaciales relativos al patrimonio cultural en el marco definido por la directiva europea INSPIRE sobre información geográfica. Abstract: This monograph presents the background, context and technical details of an Application Schema for the inclusion of cultural heritage spatial data into the INSPIRE framework. Nowadays, INSPIRE provides the most relevant framework for the dissemination and exchange of geographical data, covering many different thematic fields, particularly relevant for envi-ronmental datasets. Although cultural heritage elements are partially addressed within INSPIRE, there is no specific documentation on how these data should be considered, structured and published. This text aims to provide technical guidelines for decision makers, public administrations and the scientific community for the definition and implementation of harmonized datasets for cultural heritage, according to the interoperability principles of INSPIRE.

MedVir: an interactive representation system of multidimensional medical data applied to Traumatic Brain Injury's rehabilitation prediction

Clinicians could model the brain injury of a patient through his brain activity. However, how this model is defined and how it changes when the patient is recovering are questions yet unanswered. In this paper, the use of MedVir framework is proposed with the aim of answering these questions. Based on complex data mining techniques, this provides not only the differentiation between TBI patients and control subjects (with a 72% of accuracy using 0.632 Bootstrap validation), but also the ability to detect whether a patient may recover or not, and all of that in a quick and easy way through a visualization technique which allows interaction.

Mining mobile apps reviews to support release planning

The mobile apps market is a tremendous success, with millions of apps downloaded and used every day by users spread all around the world. For apps’ developers, having their apps published on one of the major app stores (e.g. Google Play market) is just the beginning of the apps lifecycle. Indeed, in order to successfully compete with the other apps in the market, an app has to be updated frequently by adding new attractive features and by fixing existing bugs. Clearly, any developer interested in increasing the success of her app should try to implement features desired by the app’s users and to fix bugs affecting the user experience of many of them. A precious source of information to decide how to collect users’ opinions and wishes is represented by the reviews left by users on the store from which they downloaded the app. However, to exploit such information the app’s developer should manually read each user review and verify if it contains useful information (e.g. suggestions for new features). This is something not doable if the app receives hundreds of reviews per day, as happens for the very popular apps on the market. In this work, our aim is to provide support to mobile apps developers by proposing a novel approach exploiting data mining, natural language processing, machine learning, and clustering techniques in order to classify the user reviews on the basis of the information they contain (e.g. useless, suggestion for new features, bugs reporting). Such an approach has been empirically evaluated and made available in a web-­‐based tool publicly available to all apps’ developers. The achieved results showed that the developed tool: (i) is able to correctly categorise user reviews on the basis of their content (e.g. isolating those reporting bugs) with 78% of accuracy, (ii) produces clusters of reviews (e.g. groups together reviews indicating exactly the same bug to be fixed) that are meaningful from a developer’s point-­‐of-­‐view, and (iii) is considered useful by a software company working in the mobile apps’ development market.

Is routine hospital episode data sufficient for identifying individuals with chronic kidney disease? : A comparison study with laboratory data

© The Author(s) 2014. Acknowledgements We thank the Information Services Division, Scotland, who provided the SMR01 data, and NHS Grampian, who provided the biochemistry data. We also thank the University of Aberdeen’s Data Management Team. Funding This work was supported by the Chief Scientists Office for Scotland (grant no. CZH/4/656).

A data driven study of relationships between relief and farmland abandonment in a Mediterranean region

The environmental, cultural and socio-economic causes and consequences of farmland abandonment are issues of increasing concern for researchers and policy makers. In previous studies, we proposed a new methodology for selecting the driving factors in farmland abandonment processes. Using Data Mining and GIS, it is possible to select those variables which are more significantly related to abandonment. The aim of this study is to investigate the application of the above mentioned methodology for finding relationships between relief and farmland abandonment in a Mediterranean region (SE Spain).We have taken into account up to 28 different variables in a single analysis, some of them commonly considered in land use change studies (slope, altitude, TWI, etc), but also other novel variables have been evaluated (sky view factor, terrain view factor, etc). The variable selection process provides results in line with the previous knowledge of the study area, describing some processes that are region specific (e.g. abandonment versus intensification of the agricultural activities). The European INSPIRE Directive (2007/2/EC) establishes that the digital elevation models for land surfaces should be available in all member countries, this means that the research described in this work can be extrapolated to any European country to determine whether these variables (slope, altitude, etc) are important in the process of abandonment.

Aplicación de un modelo geográfico con información climática para el cálculo del balance hídrico de la comarca de la Marina Baja (Alicante)

Este trabajo analiza las nuevas tendencias en la creación y gestión de información geográfica, para la elaboración de modelos inductivos basados exclusivamente en bases de datos geográficas. Estos modelos permiten integrar grandes volúmenes de datos de características heterogéneas, lo que supone una gran complejidad técnica y metodológica. Se propone una metodología que permite conocer detalladamente la distribución de los recursos hídricos naturales en un territorio y derivar numerosas capas de información que puedan ser incorporadas a estos modelos «ávidos de datos» (data-hungry). La zona de estudio escogida para aplicar esta metodología es la comarca de la Marina Baja (Alicante), para la que se presenta un cálculo del balance hídrico espacial mediante el uso de herramientas estadísticas, geoestadísticas y Sistemas de Información Geográfica. Finalmente, todas las capas de información generadas (84) han sido validadas y se ha comprobado que su creación admite un cierto grado de automatización que permitirá incorporarlas en análisis de Minería de Datos más amplios.

Homogeneous Spatial Units (HSU) - a Pan-European geographical basis for environmental and socio-economic modelling

The spatial data set delineates areas with similar environmental properties regarding soil, terrain morphology, climate and affiliation to the same administrative unit (NUTS3 or comparable units in size) at a minimum pixel size of 1km2. The scope of developing this data set is to provide a link between spatial environmental information (e.g. soil properties) and statistical data (e.g. crop distribution) available at administrative level. Impact assessment of agricultural management on emissions of pollutants or radiative active gases, or analysis regarding the influence of agricultural management on the supply of ecosystem services, require the proper spatial coincidence of the driving factors. The HSU data set provides e.g. the link between the agro-economic model CAPRI and biophysical assessment of environmental impacts (updating previously spatial units, Leip et al. 2008), for the analysis of policy scenarios. Recently, a statistical model to disaggregate crop information available from regional statistics to the HSU has been developed (Lamboni et al. 2016). The HSU data set consists of the spatial layers provided in vector and raster format as well as attribute tables with information on the properties of the HSU. All input data for the delineation the HSU is publicly available. For some parameters the attribute tables provide the link between the HSU data set and e.g. the soil map(s) rather than the data itself. The HSU data set is closely linked the USCIE data set.

Education and Development : A spatial-econometric approach on spatial change and permanence in southern Brazil

The Brazilian state of Paraná exhibits a violent geography of inequality and duality, hosting both the most developed city in the country, internationally recognized by its urban and environmental innovations, and southern Brazil’s most concentrated cluster of poverty and underdevelopment. Over the course of the past decades, the state underwent a major economic transformation, modernizing and increasing its industrial structure and shifting to the service sector with a larger participation of the knowledge economy. This study is concerned on the interplay between formal education and socioeconomic development during this process, and above all its spatial character. It attempts make sense of the rich literature on education and growth and/or development, discussing it through the lenses of human geography and planning. In order for the analysis to be possible, this study created a consistent database of municipal scores of education over the course of 40 years, dealing with changing census methodologies and municipal boundaries. Making use of modern exploratory spatial data analysis combined with spatial regressions, the study identifies a clustered, time-persistent interplay between education and development that is stronger for low and basic levels of education. Moreover, it provides evidence that not only education is a predictor of future development, but also that analyses of this kind must take into consideration spatial autocorrelation in order to be accurate.

Remotely sensed data for ecosystem analyses: Combining hierarchy and scene models

Remotely sensed data have been used extensively for environmental monitoring and modeling at a number of spatial scales; however, a limited range of satellite imaging systems often. constrained the scales of these analyses. A wider variety of data sets is now available, allowing image data to be selected to match the scale of environmental structure(s) or process(es) being examined. A framework is presented for use by environmental scientists and managers, enabling their spatial data collection needs to be linked to a suitable form of remotely sensed data. A six-step approach is used, combining image spatial analysis and scaling tools, within the context of hierarchy theory. The main steps involved are: (1) identification of information requirements for the monitoring or management problem; (2) development of ideal image dimensions (scene model), (3) exploratory analysis of existing remotely sensed data using scaling techniques, (4) selection and evaluation of suitable remotely sensed data based on the scene model, (5) selection of suitable spatial analytic techniques to meet information requirements, and (6) cost-benefit analysis. Results from a case study show that the framework provided an objective mechanism to identify relevant aspects of the monitoring problem and environmental characteristics for selecting remotely sensed data and analysis techniques.

Dynamic multi-resolution spatial object derivation for mobile and WWW applications

Online geographic information systems provide the means to extract a subset of desired spatial information from a larger remote repository. Data retrieved representing real-world geographic phenomena are then manipulated to suit the specific needs of an end-user. Often this extraction requires the derivation of representations of objects specific to a particular resolution or scale from a single original stored version. Currently standard spatial data handling techniques cannot support the multi-resolution representation of such features in a database. In this paper a methodology to store and retrieve versions of spatial objects at, different resolutions with respect to scale using standard database primitives and SQL is presented. The technique involves heavy fragmentation of spatial features that allows dynamic simplification into scale-specific object representations customised to the display resolution of the end-user's device. Experimental results comparing the new approach to traditional R-Tree indexing and external object simplification reveal the former performs notably better for mobile and WWW applications where client-side resources are limited and retrieved data loads are kept relatively small.

An internet GIS and spatial model to benchmark local government socio-economic performance

This paper develops an Internet geographical information system (GIS) and spatial model application that provides socio-economic information and exploratory spatial data analysis for local government authorities (LGAs) in Queensland, Australia. The application aims to improve the means by which large quantities of data may be analysed, manipulated and displayed in order to highlight trends and patterns as well as provide performance benchmarking that is readily understandable and easily accessible for decision-makers. Measures of attribute similarity and spatial proximity are combined in a clustering model with a spatial autocorrelation index for exploratory spatial data analysis to support the identification of spatial patterns of change. Analysis of socio-economic changes in Queensland is presented. The results demonstrate the usefulness and potential appeal of the Internet GIS applications as a tool to inform the process of regional analysis, planning and policy.

Fine-scale spatial genetic correlation analyses reveal strong female philopatry within a brush-tailed rock-wallaby colony in southeast Queensland

We combine spatial data on home ranges of individuals and microsatellite markers to examine patterns of fine-scale spatial genetic structure and dispersal within a brush-tailed rock-wallaby (Petrogale penicillata) colony at Hurdle Creek Valley, Queensland. Brush-tailed rock-wallabies were once abundant and widespread throughout the rocky terrain of southeastern Australia; however, populations are nearly extinct in the south of their range and in decline elsewhere. We use pairwise relatedness measures and a recent multilocus spatial autocorrelation analysis to test the hypotheses that in this species, within-colony dispersal is male-biased and that female philopatry results in spatial clusters of related females within the colony. We provide clear evidence for strong female philopatry and male-biased dispersal within this rock-wallaby colony. There was a strong, significant negative correlation between pairwise relatedness and geographical distance of individual females along only 800 m of cliff line. Spatial genetic autocorrelation analyses showed significant positive correlation for females in close proximity to each other and revealed a genetic neighbourhood size of only 600 m for females. Our study is the first to report on the fine-scale spatial genetic structure within a rock-wallaby colony and we provide the first robust evidence for strong female philopatry and spatial clustering of related females within this taxon. We discuss the ecological and conservation implications of our findings for rock-wallabies, as well as the importance of fine-scale spatial genetic patterns in studies of dispersal behaviour.