As the quantity of text documents created on the web grows the ability of experts to manually classify them has decreased. Because people need to find and organize this information, interest has grown in developing automatic means of categorizing these documents. In this effort, ontologies have been developed that capture domain specific knowledge in the form of a hierarchy of concepts. Support Vector Machines are machine learning methods that are widely used for automated document categorization. Recent studies suggest that the classification accuracy of a Support Vector Machine may be improved by using concepts defined by a domain ontology instead of using the words that appear in the document. However, such studies have not taken into account the hierarchy inherent in the relationship between concepts. The goal of this dissertation was to investigate whether the hierarchical relationships among concepts in ontologies can be exploited to improve the classification accuracy of web documents by a Support Vector Machine. Concept vectors that capture the hierarchy of domain ontologies were created and used to train a Support Vector Machine. Tests conducted using the benchmark Reuters-21578 data set indicate that the Support Vector Machines achieve higher classification accuracy when they make use of the hierarchical relationships among concepts in ontologies.

Esta investigación se centra en el problema de clasificación, por medio de confeccionar una metodología que permita detectar y modelar cambios en los patrones que definen la clasificación en el tiempo, en otras palabras, clasificación dinámica. La metodología desarrollada propone utilizar los resultados obtenidos en un periodo de tiempo para la construcción del modelo al siguiente periodo. Para ello se utilizaron dos modelos de clasificación distintos; el primero de ellos es Support Vector Machine (SVM) con el objetivo de confeccionar la metodología dinámica, que denominaremos Dynamic Support Vector Machine (D-SVM) y el segundo modelo de clasificación es Linear Penalizad SVM (LP-SVM) con la finalidad de que la metodología construida permita la selección de atributos dinámicamente. Los parámetros utilizados en el modelo de clasificación son; las ventanas de tiempo, ponderadores de relevancia, penalización de los errores y la penalización de los atributos (sólo para el modelo con selección de atributos). De los resultados obtenidos, se utiliza la ventana de tiempo que define el mejor modelo de un periodo y junto a los nuevos datos que se obtengan generan el del próximo. Esta metodología luego fue aplicada a un caso real en una institución gubernamental chilena (INDAP), en el problema de predicción de comportamiento de pago (credit scoring). Para ello se analizaron 4 instancias de tiempo con 9 atributos para el modelo sin selección de atributos y 18 atributos para el modelo con selección. Luego ambos modelos fueron comparados con uno de clasificación estática, es decir, que las 4 instancias de tiempo son unidas como si fuese una data. Los resultados obtenidos en esta aplicación son levemente superiores a la metodología estática correspondiente y en el caso de la selección de atributos el modelo utiliza una mayor cantidad. Las conclusiones de esta investigación son que presenta la ventaja de utilizar una menor cantidad de datos a los disponibles, lo que genera modelos más rápidos y que se van adaptando a los cambios de comportamiento que se producen en el tiempo, al descartar los datos más antiguos en la construcción del nuevo modelo. Con respecto al método con selección de atributos, se destaca que no utiliza un modelo exógeno para seleccionar los atributos sino que el modelo estima los atributos necesarios para cada periodo de tiempo, por lo que se tiene un modelo más estable y generalizado; además se logra obtener información de cómo la relevancia de los atributos cambia en el tiempo. Sobre los resultados se concluye que la metodología D-SVM con y sin selección de atributos es al menos tan buena como los métodos actuales de clasificación.

Made available in DSpace on 2016-01-26T18:56:03Z (GMT). No. of bitstreams: 1 Ronaldo Oikawa.pdf: 3711757 bytes, checksum: 5d13d9ef8b2d092d0bb1a6fe62a73248 (MD5) Previous issue date: 2015-03-16 / The Artificial Neural Networks are nonlinear mathematical models that resemble the human brain, and this ability to learn was applied to recognize the rain patterns in the city of Presidente Prudente, located in the region of Pontal do Paranapanema. Through these calculations, it was possible to indicate another way to rain forecast. This study used two algorithms with supervised learning, the first one the Multiple Layer Network Propagation, with 23 neurons and with one, two and three hidden layers, and the second one the Support Vector Machine (SVM) with polynomial, radial basis function and hyper tangent cores. The set analyzed covers the period from January 1996 to May 2012, collected from Weather Forecast Center and Climate Studies (CPTEC). The results showed that the atmospheric pressure, wind direction, minimum temperature and air relative humidity were the parameters more related with the rain precipitation. The SVM model with base radial function core, using sigma=0.1, showed the best results with Kappa coefficient, equal to 0.675 for first test group, equal to 0.746 to the second test group 0.746 and equal to 0.826 for the third test group. These results demonstrate the data set robustness and allowed achieve high accuracy rate in recognition of rain precipitation. / As Redes Neurais Artificiais são modelos matemáticos não lineares que se assemelham ao cérebro humano, e esta capacidade de aprender foi aplicada no reconhecimento de padrões da chuva na cidade de Presidente Prudente, localizada na região do Pontal do Paranapanema. Através desses cálculos foi possível indicar uma forma alternativa de se reconhecer o padrão da precipitação da chuva. O presente trabalho utilizou dois algoritmos com aprendizagem supervisionada, sendo o primeiro a Rede de Múltipla Camada de Retro Propagação, com 23 neurônios e com uma, duas e três camadas ocultas, já o segundo, a Máquina de Vetor de Suporte (SVM) utilizou o núcleo polinomial, função de base radial e hiper tangente. O conjunto de dados analisados compreende o período de Janeiro de 1996 até Maio de 2012, sendo obtidos do Centro de Previsão de Tempo e Estudos Climáticos (CPTEC). Os resultados demonstraram que a pressão atmosférica, direção do vento, temperatura mínima e umidade relativa do ar são os parâmetros que estão mais relacionados à precipitação da chuva. O modelo SVM, com núcleo função de base radial, utilizando o parâmetro sigma=0,1 obteve os melhores resultados, apresentando o coeficiente Kappa (resposta), igual a 0,675 para o grupo de teste um, 0,746 para o grupo de teste dois e 0,826 para o grupo de teste três. Estes resultados demonstram a robustez do conjunto de dados e permitiram atingir altos índices de acerto no reconhecimento da precipitação da chuva.

Made available in DSpace on 2016-07-18T17:46:20Z (GMT). No. of bitstreams: 1 Ronaldo Oikawa.pdf: 3711757 bytes, checksum: 5d13d9ef8b2d092d0bb1a6fe62a73248 (MD5) Previous issue date: 2015-03-16 / The Artificial Neural Networks are nonlinear mathematical models that resemble the human brain, and this ability to learn was applied to recognize the rain patterns in the city of Presidente Prudente, located in the region of Pontal do Paranapanema. Through these calculations, it was possible to indicate another way to rain forecast. This study used two algorithms with supervised learning, the first one the Multiple Layer Network Propagation, with 23 neurons and with one, two and three hidden layers, and the second one the Support Vector Machine (SVM) with polynomial, radial basis function and hyper tangent cores. The set analyzed covers the period from January 1996 to May 2012, collected from Weather Forecast Center and Climate Studies (CPTEC). The results showed that the atmospheric pressure, wind direction, minimum temperature and air relative humidity were the parameters more related with the rain precipitation. The SVM model with base radial function core, using sigma=0.1, showed the best results with Kappa coefficient, equal to 0.675 for first test group, equal to 0.746 to the second test group 0.746 and equal to 0.826 for the third test group. These results demonstrate the data set robustness and allowed achieve high accuracy rate in recognition of rain precipitation. / As Redes Neurais Artificiais são modelos matemáticos não lineares que se assemelham ao cérebro humano, e esta capacidade de aprender foi aplicada no reconhecimento de padrões da chuva na cidade de Presidente Prudente, localizada na região do Pontal do Paranapanema. Através desses cálculos foi possível indicar uma forma alternativa de se reconhecer o padrão da precipitação da chuva. O presente trabalho utilizou dois algoritmos com aprendizagem supervisionada, sendo o primeiro a Rede de Múltipla Camada de Retro Propagação, com 23 neurônios e com uma, duas e três camadas ocultas, já o segundo, a Máquina de Vetor de Suporte (SVM) utilizou o núcleo polinomial, função de base radial e hiper tangente. O conjunto de dados analisados compreende o período de Janeiro de 1996 até Maio de 2012, sendo obtidos do Centro de Previsão de Tempo e Estudos Climáticos (CPTEC). Os resultados demonstraram que a pressão atmosférica, direção do vento, temperatura mínima e umidade relativa do ar são os parâmetros que estão mais relacionados à precipitação da chuva. O modelo SVM, com núcleo função de base radial, utilizando o parâmetro sigma=0,1 obteve os melhores resultados, apresentando o coeficiente Kappa (resposta), igual a 0,675 para o grupo de teste um, 0,746 para o grupo de teste dois e 0,826 para o grupo de teste três. Estes resultados demonstram a robustez do conjunto de dados e permitiram atingir altos índices de acerto no reconhecimento da precipitação da chuva.

京都大学 / 0048 / 新制・論文博士 / 博士(医学) / 乙第13325号 / 論医博第2193号 / 新制||医||1043(附属図書館) / (主査)教授 黒田 知宏, 教授 藤渕 航, 教授 別所 和久 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

Background: In drug discovery morphological profiles can be used to identify and establish a drug's biological activity or mechanism of action. Quantitative structure-activity relationship (QSAR) is an approach that uses the chemical structures to predict properties e.g., biological activity. Support Vector Machine (SVM) is a machine learning algorithm that can be used for classification. Confidence measures as conformal predictions can be implemented on top of machine learning algorithms. There are several methods that can be applied to improve a model’s predictive performance. Aim: The aim in this project is to evaluate if ligand-based modelling can be improved by combining features from chemical structures, target predictions and morphological profiles. Method: The project was divided into three experiments. In experiment 1 five bioassay datasets were used. In experiment 2 and 3 a cell painting dataset was used that contained morphological profiles from three different classes of kinase inhibitors, and the classes were used as endpoints. Support vector machine, liblinear models were built in all three experiments. A significant level of 0.2 was set to calculate the efficiency. The mean observed fuzziness and efficiency were used as measurements to evaluate the model performance. Results: Similar trends were observed for all datasets in experiment 1. Signatures+CDK13+TP which is the most complex model obtained the lowest mean observed fuzziness in four out of five times. With a confidence level of 0.8, TP+Signatures obtained the highest efficiency. Signatures+Morphological Profiles+TP obtained the lowest mean observed fuzziness in experiment 2 and 3. Signatures obtained the highest correct single label predictions with a confidence of 80%. Discussion: Less correct single label predictions were observed for the active class in comparison to the inactive class. This could have been due to them being harder to predict. The morphological profiles did not contribute with an improvement to the models predictive performance compared to Signatures. This could be due to the lack of information obtained from the dataset. Conclusion: A combination of features from chemical structures and target predictions improved ligand-based modelling compared to models only built on one of the features. The combination of features from chemical structures and morphological profiles did not improve the ligand-based models, compared to the model only built on chemical structures. By adding features from target predictions to a model built with features from chemical structures and morphological profiles a decrease in mean observed fuzziness was obtained.

In the long term it is beneficial to a country's economy to exploit the maritime environment surrounding it responsibly. It is also beneficial to protect this environment from poaching and pollution. To achieve this the responsible parties of a country must have an awareness of what is transpiring in the maritime domain. Synthetic aperture radar can provide an image, regardless of weather or light conditions, of the ocean showing most vessels therein. To monitor the ocean, using synthetic aperture radar imagery, at the lowest cost would require large swath synthetic aperture radar imagery. There exists a trade-off between large swath imagery and the image's resolution resulting in the largest swath image having the poorest resolution. Existing research has shown that it is possible to use coarse resolution synthetic aperture radar imagery to detect vessels at sea, but little work has been done on classifying those vessels. This research aims to investigate the coarse resolution classification information gap. This is done by using a dataset of matching synthetic aperture radar and ship transponder data to train a statistical classification algorithm in order to classify or estimate the length of vessels based on features extracted from their synthetic aperture radar image. The results of this research show that coarse resolution (approximately 40 m per pixel) synthetic aperture radar imagery is able to estimate vessel size for larger classes and provides insight on which vessel classes would require finer resolutions in order to be detected and classified reliably. The range of smaller vessel classes is usually limited to ports and fishing zones. These zones can be mapped using historical vessel transponder data and so a dedicated surveillance campaign can be optimised to use higher resolution products in these areas. The size estimation from the machine learning algorithm performs better than current techniques. / Dissertation (MEng)--University of Pretoria, 2017. / Electrical, Electronic and Computer Engineering / MEng / Unrestricted

Authorship identification is the field of classifying a given text by its author based on the assumption that authors exhibit unique writing styles. This thesis investigates the semantic shortcomings of the vector space model by constructing a semantic kernel created from WordNet which is evaluated on the problem of authorship attribution. A multiclass SVM classifier is constructed using the one-versus-all strategy and evaluated in terms of precision, recall, accuracy and F1 scores. Results show that the use of the semantic scores from WordNet degrades the performance compared to using a linear kernel. Experiments are run to identify the best feature engineering configurations, showing that removing stopwords has a positive effect on the financial dataset Reuters while the Kaggle dataset consisting of short extracts of horror stories benefit from keeping the stopwords.

In this paper, Natural Language Processing and classification algorithms were used to create a program that automatically can tag different errands that are connected to Fortnox (an IT company based in Växjö) support service. Controlled experiments were conducted to find the best classification algorithm together with different Bag-of-Word pre-processing algorithms to find what was best suited for this problem. All data were provided by Fortnox and were manually labeled with tags connected to it as training and test data. The result of the final algorithm was 69.15% correctly/accurately predicted errands using all original data. When looking at the data that were incorrectly predicted a pattern was noticed where many errands have identical text attached to them. By removing the majority of these errands, the result was increased to 94.08%.

Head injury is common in many different sports and elsewhere, and is often associated with differentdifficulties. One major problem is to identify and value the injury or the severity. Sometimes there is no sign of head injury, but a serious neck distortion has occurred, causing similar symptoms as head injuries e.g. concussion or mild TBI (traumatic brain injury). This study investigated whether direct and indirect measurements of head kinematics, combined with machine learning and 3D visualization can be used to identify head injury and value the injury. Injury statistics have found that many severe head injuries are caused by oblique impacts. An oblique impact will give rise to both linear and rotational kinematics. Since the human brain is very sensitive to rotational kinematics, many violent rotations of the head can results in large shear strains in the brain. This is when white matter and white matter connections are disrupted in the brain from acceleration and deceleration, or rotational acceleration kinematics which in turn will cause traumatic brain injuries as e.g. diffuse axonal injury (DAI). Lately there has been many studies in this field using different types of new technologies, but the most prevalent is the rise of wearable sensors that have become smaller, faster and more energy efficient where they have been integrated into mouthguards and inertial measurement units (IMUs) the size of a sim-card that measures and reports a body's specific force. It has been shown that a 6-axis IMU (3-axis rotational- and 3-axis acceleration measurements) may improve head injury prediction but more data is needed to confirm with existing head injury criterions and new criterions needs to be developed, that considers directional sensitivity. Today, IMUs are typically used in self-driving cars, aircrafts, spacecrafts, satellites etc. As of today, more and more studies have evaluated and utilized IMUs in new uncharted fields have shown promises, especially in sports, and in the neuroscience and medical field. This study proposed a method to 3D visualize head kinematics during the event of a possible head injury to indirectly identify and value the injury, by medical professionals, as well as, a direct method to identify and also value the severity of head injury with machine learning. An erroneous data collection process of reconstructed head impacts and non-head impacts have been recorded using an open-source 9-axis IMU sensor and a proprietary 6-axis IMU sensor. To value the head injury or the severity, existing head injury criterions as the Abbreviated Injury Scale (AIS), Head Injury Criterion (HIC), Head Impact Power (HIP), Severity Index (SI) and Generalized Acceleration Model for Brain Injury Threshold (GAMBIT) have been introduced. To detect head impact including the severity and non-head impact, a Random Forests (RF) classifier and Support Vector Machine (SVM) classifiers with linear- and radial basis function have been proposed, the prediction results have been promising.