Many protein structure prediction programs exist and they can efficiently generate a number of protein models of a varying quality. One of the problems is that it is difficult to know which model is the best one for a given target sequence. Selecting the best model is one of the major tasks of Model Quality Assessment Programs (MQAPs). These programs are able to predict model accuracy before the native structure is determined. The accuracy estimation can be divided into two parts: global (the whole model accuracy) and local (the accuracy of each residue). ProQ2 is one of the most successful MQAPs for prediction of both local and global model accuracy and is based on a Machine Learning approach. In this thesis, I present my own contribution to Model Quality Assessment (MQA) and the newest developments of ProQ program series. Firstly, I describe a new ProQ2 implementation in the protein modelling software package Rosetta. This new implementation allows use of ProQ2 as a scoring function for conformational sampling inside Rosetta, which was not possible before. Moreover, I present two new methods, ProQ3 and ProQ3D that both outperform their predecessor. ProQ3 introduces new training features that are calculated from Rosetta energy functions and ProQ3D introduces a new machine learning approach based on deep learning. ProQ3 program participated in the 12th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP12) and was one of the best methods in the MQA category. Finally, an important issue in model quality assessment is how to select a target function that the predictor is trying to learn. In the fourth manuscript, I show that MQA results can be improved by selecting a contact-based target function instead of more conventional superposition based functions. / <p>At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 3: Manuscript.</p>

Orientador: Ronei Jesus Popi / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Química / Made available in DSpace on 2018-08-18T17:27:36Z (GMT). No. of bitstreams: 1 Maretto_DaniloAlthmann_D.pdf: 2617064 bytes, checksum: 1ebea2b6ab73ef552155cd9b79b6fd1b (MD5) Previous issue date: 2011 / Resumo: O objetivo desta tese de doutorado foi de utilizar o algoritmo Máquinas de Vetores de Suporte (SVM) em problemas de classificação e calibração, onde algoritmos mais tradicionais (SIMCA e PLS, respectivamente) encontram problemas. Foram realizadas quatro aplicações utilizando dados de espectroscopia no infravermelho. Na primeira o SVM se mostrou ser uma ferramenta mais indicada para a determinação de Carbono e Nitrogênio em solo por NIR, quando estes elementos estão em solos sem que se saiba se há ou não a presença do mineral gipsita, obtendo concentrações desses elementos com erros consideravelmente menores do que a previsão feita pelo PLS. Na determinação da concentração de um mineral em polímero por NIR, que foi a segunda aplicação, o PLS conseguiu previsões com erros aceitáveis, entretanto, através da análise do teste F e o gráfico de erros absolutos das previsões, foi possível concluir que o modelo SVM conseguiu chegar a um modelo mais ajustado. Na terceira aplicação, que consistiu na classificação de bactérias quanto às condições de crescimento (temperaturas 30 ou 40°C e na presença ou ausência de fosfato) por MIR, o SIMCA não foi capaz de classificar corretamente a grande maioria das amostras enquanto o SVM produziu apenas uma previsão errada. E por fim, na última aplicação, que foi a diferenciação de nódulos cirróticos e de hepatocarcinoma por microespectroscopia MIR, a taxa das previsões corretas para os conjuntos de validação do SVM foram maiores do que do SIMCA. Nas quatro aplicações o SVM produziu resultados melhores do que o SIMCA e o PLS, mostrando que pode ser uma alternativa aos métodos mais tradicionais de classificação e calibração multivariada / Abstract: The objective of this thesis was to use the algorithm Support Vector Machines (SVM) in problems of classification and calibration, where more traditional algorithms (SIMCA and PLS, respectively) present problems. Four applications were developed using data for infrared spectra. In the first one, the SVM proved to be a most suitable tool for determination of carbon and nitrogen in soil by NIR, when these elements are in soils without knowledge whether or not the presence of the gypsum mineral, obtaining concentrations of these elements with errors considerably smaller than the estimated by the PLS. In the determination of the concentration of a mineral in a polymer by NIR, which was the second application, the PLS presented predictions with acceptable errors, however, by examining the F test and observing absolute errors of predictions, it was concluded that the SVM was able to reach a more adjusted model. In the third application, classification of bacteria on the different growth conditions (temperatures 30 or 40 ° C and in the presence or absence of phosphate) by MIR, the SIMCA was not able to correctly classify the majority of the samples while the SVM produced only one false prediction. Finally, in the last application, which was the differentiation of cirrhotic nodules and Hepatocellular carcinoma by infrared microspectroscopy, the rate of correct predictions for the validation of sets of SVM was higher than the SIMCA. In the four applications SVM produced better results than SIMCA and PLS, showing that it can be an alternative to the traditional algorithms for classification and multivariate calibration / Doutorado / Quimica Analitica / Doutor em Ciências

This thesis deals with the design and implementation of a tool for generating drums triggers from a downmix record. The work describes the preprocessing of the input audio signal and methods for the classification of strokes. The drum classification is based on the similarity of the signals in the frequency domain. Principal component analysis (PCA) was used to reduce the number of dimensions and to find the characteristic properties of the input data. The method support vector machine (SVM) was used to classify the data into individual classes representing parts of the drum kit. The software was programmed in Matlab. The classification model was trained on a set of 728 drum samples for seven categories (kick, snare, hi-hat, crash, ride, kick + hi-hat, snare + hi-hat). The success of the system in the classification is 75 %.

No description available.

Object recognition is one of many tasks in which the computer is still behind the human. Therefore, development in this area takes inspiration from nature and especially from the function of the human brain. This work focuses on object recognition based on extracting relevant information from images, features. Features are obtained in a similar way as the human brain processes visual stimuli. Subsequently, these features are used to train classifiers for object recognition (e.g. SVM, k-NN, ANN). This work examines the feature extraction stage. Its aim is to improve the feature extraction and thereby increase performance of object recognition by computer.

The document is discussing the issue of the computer vision with ability to character recignition in the image. Wavelet transform is used for preprocessing the image. Pixel energy feature is firstly used for searchich candidate text pixels. Density region growing method is then used to collect candidate pixels to the separate regions, which will be candidate text regions. Several of the features are calculated over the regions and the SVM classifier is used to derive, if the region is really a text region or not.

The aim of this diploma thesis is to get acquainted with the properties of image data and the principle of their capture. Literary research on methods of image segmentation in the area of cardiac tissue imaging and, last but not least, efforts to find methods for classification of dead cardiomyocytes and analysis of their viability. Dead cardiomyocytes were analyzed for their shape and similarity to the template created as a mean of dead cells. Another approach was the application of the method based on local binary characters and the computation of symptoms from a simple and associated histogram.

Background. Since historical handwritten documents have played important roles in promoting the development of human civilization, many of them have been preserved through digital versions for more scientific researches. However, various degradations always exist in these documents, which could interfere in normal reading. But, binarized versions can keep meaningful contents without degradations from original document images. Document image binarization always works as a pre-processing step before complex document analysis and recognition. It aims to extract texts from a document image. A desirable binarization performance can promote subsequent processing steps positively. For getting better performance for document image binarization, efficient binarization methods are needed. In recent years, machine learning centered on deep learning has gathered substantial attention in document image binarization, for example, Convolutional Neural Networks (CNNs) are widely applied in document image binarization because of the powerful ability of feature extraction and classification. Meanwhile, Support Vector Machine (SVM) is also used in image binarization. Its objective is to build an optimal hyperplane that could maximize the margin between negative samples and positive samples, which can separate the foreground pixels and the background pixels of the image distinctly. Objectives. This thesis aims to explore how the CNN based process of deep convolutional feature extraction and an SVM classifier can be integrated well to binarize handwritten document images, and how the results are, compared with some state-of-the-art document binarization methods. Methods. To investigate the effect of the proposed method on document image binarization, it is implemented and trained. In the architecture, CNN is used to extract features from input images, afterwards these features are fed into SVM for classification. The model is trained and tested with six different datasets. Then, there is a performance comparison between the proposed model and other binarization methods, including some state-of-the-art methods on other three different datasets. Results. The performance results indicate that the proposed model not only can work well but also perform better than some other novel handwritten document binarization method. Especially, evaluation of the results on DIBCO 2013 dataset indicates that our method fully outperforms other chosen binarization methods on all the four evaluation metrics. Besides, it also has the ability to deal with some degradations, which demonstrates its generalization and learning ability are excellent. When a new kind of degradation appears, the proposed method can address it properly even though it never appears in the training datasets. Conclusions. This thesis concludes that the CNN based component and SVM can be combined together for handwritten document binarization. Additionally, in certain datasets, it outperforms some other state-of-the-art binarization methods. Meanwhile, its generalization and learning ability is outstanding when dealing with some degradations.

No description available.

Stock Forecasting is commonly used in different forms everyday in order to predict stock prices. Sentiment Analysis (SA), Machine Learning (ML) and Data Mining (DM) are techniques that have recently become popular in analyzing public emotion in order to predict future stock prices. The algorithms need data in big sets to detect patterns, and the data has been collected through a live stream for the tweet data, together with web scraping for the stock data. This study examined how three organization's stocks correlate with the public opinion of them on the social networking platform, Twitter. Implementing various machine learning and classification models such as the Artificial Neural Network we successfully implemented a company-specific model capable of predicting stock price movement with 80% accuracy.