Text classification using compression-based dissimilarity measures

Arguably, the most difficult task in text classification is to choose an appropriate set of features that allows machine learning algorithms to provide accurate classification. Most state-of-the-art techniques for this task involve careful feature engineering and a pre-processing stage, which may be too expensive in the emerging context of massive collections of electronic texts. In this paper, we propose efficient methods for text classification based on information-theoretic dissimilarity measures, which are used to define dissimilarity-based representations. These methods dispense with any feature design or engineering, by mapping texts into a feature space using universal dissimilarity measures; in this space, classical classifiers (e.g. nearest neighbor or support vector machines) can then be used. The reported experimental evaluation of the proposed methods, on sentiment polarity analysis and authorship attribution problems, reveals that it approximates, sometimes even outperforms previous state-of-the-art techniques, despite being much simpler, in the sense that they do not require any text pre-processing or feature engineering.

Thermo-compression molding of chitosan with a deep eutectic mixture for biofilms development

Supplementary information available at: http://www.rsc.org/suppdata/c5/gc/c5gc02231b/c5gc02231b1.pdf

Compression-based Image Registration

Image registration is an important component of image analysis used to align two or more images. In this paper, we present a new framework for image registration based on compression. The basic idea underlying our approach is the conjecture that two images are correctly registered when we can maximally compress one image given the information in the other. The contribution of this paper is twofold. First, we show that the image registration process can be dealt with from the perspective of a compression problem. Second, we demonstrate that the similarity metric, introduced by Li et al., performs well in image registration. Two different versions of the similarity metric have been used: the Kolmogorov version, computed using standard real-world compressors, and the Shannon version, calculated from an estimation of the entropy rate of the images

Analysis of fatty acids in ecstasy tablets

Fatty acids are the basis of so-called stearates which are frequently used as lubricants in the production of ecstasy tablets. Being a product added at the initial tablet production step its composition does not change once the compression is performed. The analysis of fatty acids can therefore provide useful information for a drug intelligence purpose. In this context an appropriate analytical method was developed to improve results already obtained by routine analyses. Considering the small quantity of such fatty acids in ecstasy tablets (not, vert, similar3%) the research focussed on their extraction and concentration. Two different procedures were tested: (1) liquid/liquid extraction using dichloromethane followed by derivatisation and (2) in situ transesterification using bortrifluoride. Analyses were performed by GC-MS. The two procedures were optimized and applied to eight ecstasy seizures, in order to choose one of the procedures for its application to a large ecstasy sample set. They were compared by considering the number of peaks detected and sample amount needed, reproducibility and other technical aspects.

Effect of compression wood on surface roughness and surface absorption of medium density fiberboard

Abstract

Dissecting the compression of mortality in Switzerland, 1876-2005

This paper aims to examine changes in common longevity and variability of the adult life span, and attempts to answer whether or not the compression of mortality continues in Switzerland in the years 1876-2005. The results show that the negative relationships between the large increase in the adult modal age at death, observed at least from the 1920s, and the decrease in the standard deviation of the ages at deaths occurring above it, illustrate a significant compression of adult mortality. Typical adult longevity increased by about 10% during the last fifty years in Switzerland, and adult heterogeneity in the age at death decreased in the same proportion. This analysis has not found any evidence suggesting that we are approaching longevity limits in term of modal or even maximum life spans. It ascertains a slowdown in the reduction of adult heterogeneity in longevity, already observed in Japan and other low mortality countries.

Patterns of neurovascular compression in patients with classic trigeminal neuralgia: A high-resolution MRI-based study.

PURPOSE: To describe the anatomical characteristics and patterns of neurovascular compression in patients suffering classic trigeminal neuralgia (CTN), using high-resolution magnetic resonance imaging (MRI). MATERIALS AND METHODS: The analysis of the anatomy of the trigeminal nerve, brain stem and the vascular structures related to this nerve was made in 100 consecutive patients treated with a Gamma Knife radiosurgery for CTN between December 1999 and September 2004. MRI studies (T1, T1 enhanced and T2-SPIR) with axial, coronal and sagital simultaneous visualization were dynamically assessed using the software GammaPlan?. Three-dimensional reconstructions were also developed in some representative cases. RESULTS: In 93 patients (93%), there were one or several vascular structures in contact, either, with the trigeminal nerve, or close to its origin in the pons. The superior cerebellar artery was involved in 71 cases (76%). Other vessels identified were the antero-inferior cerebellar artery, the basilar artery, the vertebral artery, and some venous structures. Vascular compression was found anywhere along the trigeminal nerve. The mean distance between the nerve compression and the origin of the nerve in the brainstem was 3.76±2.9mm (range 0-9.8mm). In 39 patients (42%), the vascular compression was located proximally and in 42 (45%) the compression was located distally. Nerve dislocation or distortion by the vessel was observed in 30 cases (32%). CONCLUSIONS: The findings of this study are similar to those reported in surgical and autopsy series. This non-invasive MRI-based approach could be useful for diagnostic and therapeutic decisions in CTN, and it could help to understand its pathogenesis.

Lossless compression of hyperspectral images

Technological progress has made a huge amount of data available at increasing spatial and spectral resolutions. Therefore, the compression of hyperspectral data is an area of active research. In somefields, the original quality of a hyperspectral image cannot be compromised andin these cases, lossless compression is mandatory. The main goal of this thesisis to provide improved methods for the lossless compression of hyperspectral images. Both prediction- and transform-based methods are studied. Two kinds of prediction based methods are being studied. In the first method the spectra of a hyperspectral image are first clustered and and an optimized linear predictor is calculated for each cluster. In the second prediction method linear prediction coefficients are not fixed but are recalculated for each pixel. A parallel implementation of the above-mentioned linear prediction method is also presented. Also,two transform-based methods are being presented. Vector Quantization (VQ) was used together with a new coding of the residual image. In addition we have developed a new back end for a compression method utilizing Principal Component Analysis (PCA) and Integer Wavelet Transform (IWT). The performance of the compressionmethods are compared to that of other compression methods. The results show that the proposed linear prediction methods outperform the previous methods. In addition, a novel fast exact nearest-neighbor search method is developed. The search method is used to speed up the Linde-Buzo-Gray (LBG) clustering method.

Aorto-bronchial and aorto-pulmonary fistulation after thoracic endovascular aortic repair: an analysis from the European Registry of Endovascular Aortic Repair Complications.

OBJECTIVES: To learn upon incidence, underlying mechanisms and effectiveness of treatment strategies in patients with central airway and pulmonary parenchymal aorto-bronchial fistulation after thoracic endovascular aortic repair (TEVAR). METHODS: Analysis of an international multicentre registry (European Registry of Endovascular Aortic Repair Complications) between 2001 and 2012 with a total caseload of 4680 TEVAR procedures (14 centres). RESULTS: Twenty-six patients with a median age of 70 years (interquartile range: 60-77) (35% female) were identified. The incidence of either central airway (aorto-bronchial) or pulmonary parenchymal (aorto-pulmonary) fistulation (ABPF) in the entire cohort after TEVAR in the study period was 0.56% (central airway 58%, peripheral parenchymal 42%). Atherosclerotic aneurysm formation was the leading indication for TEVAR in 15 patients (58%). The incidence of primary endoleaks after initial TEVAR was n = 10 (38%), of these 80% were either type I or type III endoleaks. Fourteen patients (54%) developed central left bronchial tree lesions, 11 patients (42%) pulmonary parenchymal lesions and 1 patient (4%) developed a tracheal lesion. The recognized mechanism of ABPF was external compression of the bronchial tree in 13 patients (50%), the majority being due to endoleak formation, further ischaemia due to extensive coverage of bronchial feeding arteries in 3 patients (12%). Inflammation and graft erosion accounted for 4 patients (30%) each. Cumulative survival during the entire study period was 39%. Among deaths, 71% were attributed to ABPF. There was no difference in survival in patients having either central airway or pulmonary parenchymal ABPF (33 vs 45%, log-rank P = 0.55). Survival with a radical surgical approach was significantly better when compared with any other treatment strategy in terms of overall survival (63 vs 32% and 63 vs 21% at 1 and 2 years, respectively), as well as in terms of fistula-related survival (63 vs 43% and 63 vs 43% at 1 and 2 years, respectively). CONCLUSIONS: ABPF is a rare but highly lethal complication after TEVAR. The leading mechanism behind ABPF seems to be a continuing external compression of either the bronchial tree or left upper lobe parenchyma. In this setting, persisting or newly developing endoleak formation seems to play a crucial role. Prognosis does not differ in patients with central airway or pulmonary parenchymal fistulation. Radical bronchial or pulmonary parenchymal repair in combination with stent graft removal and aortic reconstruction seems to be the most durable treatment strategy.

Sentence Compression with Supervised Machine Learning

The subject of the thesis is automatic sentence compression with machine learning, so that the compressed sentences remain both grammatical and retain their essential meaning. There are multiple possible uses for the compression of natural language sentences. In this thesis the focus is generation of television program subtitles, which often are compressed version of the original script of the program. The main part of the thesis consists of machine learning experiments for automatic sentence compression using different approaches to the problem. The machine learning methods used for this work are linear-chain conditional random fields and support vector machines. Also we take a look which automatic text analysis methods provide useful features for the task. The data used for machine learning is supplied by Lingsoft Inc. and consists of subtitles in both compressed an uncompressed form. The models are compared to a baseline system and comparisons are made both automatically and also using human evaluation, because of the potentially subjective nature of the output. The best result is achieved using a CRF - sequence classification using a rich feature set. All text analysis methods help classification and most useful method is morphological analysis. Tutkielman aihe on suomenkielisten lauseiden automaattinen tiivistäminen koneellisesti, niin että lyhennetyt lauseet säilyttävät olennaisen informaationsa ja pysyvät kieliopillisina. Luonnollisen kielen lauseiden tiivistämiselle on monta käyttötarkoitusta, mutta tässä tutkielmassa aihetta lähestytään television ohjelmien tekstittämisen kautta, johon käytännössä kuuluu alkuperäisen tekstin lyhentäminen televisioruudulle paremmin sopivaksi. Tutkielmassa kokeillaan erilaisia koneoppimismenetelmiä tekstin automaatiseen lyhentämiseen ja tarkastellaan miten hyvin erilaiset luonnollisen kielen analyysimenetelmät tuottavat informaatiota, joka auttaa näitä menetelmiä lyhentämään lauseita. Lisäksi tarkastellaan minkälainen lähestymistapa tuottaa parhaan lopputuloksen. Käytetyt koneoppimismenetelmät ovat tukivektorikone ja lineaarisen sekvenssin mallinen CRF. Koneoppimisen tukena käytetään tekstityksiä niiden eri käsittelyvaiheissa, jotka on saatu Lingsoft OY:ltä. Luotuja malleja vertaillaan Lopulta mallien lopputuloksia evaluoidaan automaattisesti ja koska teksti lopputuksena on jossain määrin subjektiivinen myös ihmisarviointiin perustuen. Vertailukohtana toimii kirjallisuudesta poimittu menetelmä. Tutkielman tuloksena paras lopputulos saadaan aikaan käyttäen CRF sekvenssi-luokittelijaa laajalla piirrejoukolla. Kaikki kokeillut teksin analyysimenetelmät auttavat luokittelussa, joista tärkeimmän panoksen antaa morfologinen analyysi.

Nonlinear Analysis of Shear Dominant Prestressed Concrete Beams using ANSYS

This study reports the details of the finite element analysis of eleven shear critical partially prestressed concrete T-beams having steel fibers over partial or full depth. Prestressed T-beams having a shear span to depth ratio of 2.65 and 1.59 that failed in shear have been analyzed using the ‘ANSYS’ program. The ‘ANSYS’ model accounts for the nonlinearity, such as, bond-slip of longitudinal reinforcement, postcracking tensile stiffness of the concrete, stress transfer across the cracked blocks of the concrete and load sustenance through the bridging action of steel fibers at crack interface. The concrete is modeled using ‘SOLID65’- eight-node brick element, which is capable of simulating the cracking and crushing behavior of brittle materials. The reinforcement such as deformed bars, prestressing wires and steel fibers have been modeled discretely using ‘LINK8’ – 3D spar element. The slip between the reinforcement (rebars, fibers) and the concrete has been modeled using a ‘COMBIN39’- nonlinear spring element connecting the nodes of the ‘LINK8’ element representing the reinforcement and nodes of the ‘SOLID65’ elements representing the concrete. The ‘ANSYS’ model correctly predicted the diagonal tension failure and shear compression failure of prestressed concrete beams observed in the experiment. The capability of the model to capture the critical crack regions, loads and deflections for various types of shear failures in prestressed concrete beam has been illustrated.

Example Based Image Analysis and Synthesis

Image analysis and graphics synthesis can be achieved with learning techniques using directly image examples without physically-based, 3D models. In our technique: -- the mapping from novel images to a vector of "pose" and "expression" parameters can be learned from a small set of example images using a function approximation technique that we call an analysis network; -- the inverse mapping from input "pose" and "expression" parameters to output images can be synthesized from a small set of example images and used to produce new images using a similar synthesis network. The techniques described here have several applications in computer graphics, special effects, interactive multimedia and very low bandwidth teleconferencing.

Sparse Correlation Kernel Analysis and Reconstruction

This paper presents a new paradigm for signal reconstruction and superresolution, Correlation Kernel Analysis (CKA), that is based on the selection of a sparse set of bases from a large dictionary of class- specific basis functions. The basis functions that we use are the correlation functions of the class of signals we are analyzing. To choose the appropriate features from this large dictionary, we use Support Vector Machine (SVM) regression and compare this to traditional Principal Component Analysis (PCA) for the tasks of signal reconstruction, superresolution, and compression. The testbed we use in this paper is a set of images of pedestrians. This paper also presents results of experiments in which we use a dictionary of multiscale basis functions and then use Basis Pursuit De-Noising to obtain a sparse, multiscale approximation of a signal. The results are analyzed and we conclude that 1) when used with a sparse representation technique, the correlation function is an effective kernel for image reconstruction and superresolution, 2) for image compression, PCA and SVM have different tradeoffs, depending on the particular metric that is used to evaluate the results, 3) in sparse representation techniques, L_1 is not a good proxy for the true measure of sparsity, L_0, and 4) the L_epsilon norm may be a better error metric for image reconstruction and compression than the L_2 norm, though the exact psychophysical metric should take into account high order structure in images.

Compression-based Image Registration

Image registration is an important component of image analysis used to align two or more images. In this paper, we present a new framework for image registration based on compression. The basic idea underlying our approach is the conjecture that two images are correctly registered when we can maximally compress one image given the information in the other. The contribution of this paper is twofold. First, we show that the image registration process can be dealt with from the perspective of a compression problem. Second, we demonstrate that the similarity metric, introduced by Li et al., performs well in image registration. Two different versions of the similarity metric have been used: the Kolmogorov version, computed using standard real-world compressors, and the Shannon version, calculated from an estimation of the entropy rate of the images