Learning higher-level features with convolutional restricted Boltzmann machines for sentiment analysis

In recent years, learning word vector representations has attracted much interest in Natural Language Processing. Word representations or embeddings learned using unsupervised methods help addressing the problem of traditional bag-of-word approaches which fail to capture contextual semantics. In this paper we go beyond the vector representations at the word level and propose a novel framework that learns higher-level feature representations of n-grams, phrases and sentences using a deep neural network built from stacked Convolutional Restricted Boltzmann Machines (CRBMs). These representations have been shown to map syntactically and semantically related n-grams to closeby locations in the hidden feature space. We have experimented to additionally incorporate these higher-level features into supervised classifier training for two sentiment analysis tasks: subjectivity classification and sentiment classification. Our results have demonstrated the success of our proposed framework with 4% improvement in accuracy observed for subjectivity classification and improved the results achieved for sentiment classification over models trained without our higher level features.

Feminist poststructural discourse analysis

Feminist poststructuralist discourse analysis (FPDA) is an approach to analyzing spoken interactions that focuses on the ways in which speakers negotiate their subject positions within competing and interwoven discourses. This article identifies the theoretical background to FPDA, its key principles, its distinctiveness from other approaches such as critical discourse analysis, and outlines some of the main directions in current research.

Individual identification via electrocardiogram analysis

Background: During last decade the use of ECG recordings in biometric recognition studies has increased. ECG characteristics made it suitable for subject identification: it is unique, present in all living individuals, and hard to forge. However, in spite of the great number of approaches found in literature, no agreement exists on the most appropriate methodology. This study aimed at providing a survey of the techniques used so far in ECG-based human identification. Specifically, a pattern recognition perspective is here proposed providing a unifying framework to appreciate previous studies and, hopefully, guide future research. Methods: We searched for papers on the subject from the earliest available date using relevant electronic databases (Medline, IEEEXplore, Scopus, and Web of Knowledge). The following terms were used in different combinations: electrocardiogram, ECG, human identification, biometric, authentication and individual variability. The electronic sources were last searched on 1st March 2015. In our selection we included published research on peer-reviewed journals, books chapters and conferences proceedings. The search was performed for English language documents. Results: 100 pertinent papers were found. Number of subjects involved in the journal studies ranges from 10 to 502, age from 16 to 86, male and female subjects are generally present. Number of analysed leads varies as well as the recording conditions. Identification performance differs widely as well as verification rate. Many studies refer to publicly available databases (Physionet ECG databases repository) while others rely on proprietary recordings making difficult them to compare. As a measure of overall accuracy we computed a weighted average of the identification rate and equal error rate in authentication scenarios. Identification rate resulted equal to 94.95 % while the equal error rate equal to 0.92 %. Conclusions: Biometric recognition is a mature field of research. Nevertheless, the use of physiological signals features, such as the ECG traits, needs further improvements. ECG features have the potential to be used in daily activities such as access control and patient handling as well as in wearable electronics applications. However, some barriers still limit its growth. Further analysis should be addressed on the use of single lead recordings and the study of features which are not dependent on the recording sites (e.g. fingers, hand palms). Moreover, it is expected that new techniques will be developed using fiducials and non-fiducial based features in order to catch the best of both approaches. ECG recognition in pathological subjects is also worth of additional investigations.

English in a global voluntary work context:a case study of spoken interaction and its implications for language pedagogy

This study analyses a sample of spoken interaction between a Japanese volunteer working for JICA (Japan International Co-operation Agency) and one of her co-workers in Jamaica. Details of the research context are provided, followed by a theoretical grounding of the project, which relates to publications in English as a Lingua Franca and related fields. In terms of methodology and epistemology, the research aligns with discourse analysis, specifically linguistic ethnography and interactional sociolinguistics. After presenting an an analysis of the spoken interaction based on these approaches, the resulting implications for language pedagogy are considered. This includes recommendations for specific aspects of language teaching and testing practice based on the research findings, which could be incorporated into a needs-driven localized pedagogy for future Japanese volunteers. These findings also carry significant implications for other contexts of language education, not only in terms of specific pedagogical practices but also regarding broader conceptions of language and communication.

Psychological influence of the media on patients commencing oral anticoagulation in atrial fibrillation:a qualitative analysis

Background: Oral anticoagulation (OAC) reduces stroke risk in patients with atrial fibrillation (AF); however it is still underutilized and sometimes refused by patients. Two inter-related studies were undertaken to understand the experiences and what influences this un- derutilisation of warfarin treatment in AF patients. These studies explored physician and patient experiences of AF and OAC treatment. The paper focuses on specific sub-themes from the study that explored patients’ experiences will be discussed. Aim: The study in question aimed to explore the experiences which influence patients’ decisions to accept, decline or discontinue OAC. Methods: Semi-structured individual interviews with patients were con- ducted. Three sub-groups of patients (n = 11) diagnosed with AF were interviewed; those who accepted, refused, and who discontinued war- farin. Interpretative phenomenological analysis (IPA) was used to examine the data. IPA is a qualitative method that focuses on how participants make sense of an experiences phenomenon Results: Three over-arching themes comprised patients’ experiences: (i)the initial consultation, (ii) life after the consultation, and (iii) patients’reflections. In the last theme, patients reflected on their perceptions ofaspirin and warfarin. Aspirin was perceived as a natural wonder-drugwhile warfarin was perceived as a dangerous drug usually given to peo-ple at the end of their life. Interestingly they perceive both drugs as‘old’. However, for aspirin it had a positive association, old meaningtried and tested. While for warfarin, old meant ‘has been around fortoo long’.Conclusion: Media had an important role in how patients’ perceptionsof these two drugs were influenced. Literature shows that framingtechniques, i.e. using certain words or phrases such as ‘rat poison’, areprocesses adopted by media to alter medical knowledge into lay per-son’s language. Patients in turn form negative cognitive schemas,between the word ‘poison’ and warfarin, leading to the negative per-ception of warfarin which could influence non-adherence to treatment.This qualitative research highlighted the potential influences of themedia on AF patient perceptions commencing OAC treatment. Theassociation between media stimuli and patient perceptions on OACshould be further explored. The influential power of lay-media couldalso be instrumental in disseminating appropriate educational materialto the public

On stopwords, filtering and data sparsity for sentiment analysis of Twitter

Sentiment classification over Twitter is usually affected by the noisy nature (abbreviations, irregular forms) of tweets data. A popular procedure to reduce the noise of textual data is to remove stopwords by using pre-compiled stopword lists or more sophisticated methods for dynamic stopword identification. However, the effectiveness of removing stopwords in the context of Twitter sentiment classification has been debated in the last few years. In this paper we investigate whether removing stopwords helps or hampers the effectiveness of Twitter sentiment classification methods. To this end, we apply six different stopword identification methods to Twitter data from six different datasets and observe how removing stopwords affects two well-known supervised sentiment classification methods. We assess the impact of removing stopwords by observing fluctuations on the level of data sparsity, the size of the classifier's feature space and its classification performance. Our results show that using pre-compiled lists of stopwords negatively impacts the performance of Twitter sentiment classification approaches. On the other hand, the dynamic generation of stopword lists, by removing those infrequent terms appearing only once in the corpus, appears to be the optimal method to maintaining a high classification performance while reducing the data sparsity and substantially shrinking the feature space

Something to talk about: Content and Language Integrated Learning in Modern Languages in British Higher Education. A case study of German at Aston University

This thesis investigates Content and Language Integrated Learning (CLIL) in German undergraduate programmes in the UK. At its core is a study of how one German department integrates the teaching of language and content in its undergraduate programmes and how instructors and students experience this approach. This micro-context is embedded in the wider macro-context of UK Higher Education and subject to outside forces - be they political, economic, socio-cultural - whose effects will manifest in more or less obvious ways. Data was collected via an online survey of Heads of German at British universities to determine the status quo of CLIL in UK Higher Education and to investigate how certain institutional parameters determine the introduction of CLIL in Higher Education. This project employs a mixed-method case study approach and is based on student questionnaires and semi-structured interview with German teaching staff. The study brings to light a number of significant aspects. For example, contrary to popular belief, content provision in the L2 is rather common at British universities, which is currently not reflected in the research. Student data indicates that German students perceive clear advantages in the university’s approach to CLIL. They consider German-taught content classes challenging yet beneficial for their language development. Staff interviews have yielded intriguing information about perceived advantages and disadvantages of CLIL, about its implications for classroom practice, and about instructors’ attitude towards teacher training, which echo findings from similar investigations in European contexts. Finally, the results of the macro-analysis and the case study are compared and contrasted with findings from European research on ICLHE/CLIL to determine differences and similarities with the British context, a set of recommendations is made regarding CLIL practice at the case study institution, and some implications these indings may have for the future of CLIL in British higher education are discussed.

Analysis and interpretation of serial position data

The representation of serial position in sequences is an important topic in a variety of cognitive areas including the domains of language, memory, and motor control. In the neuropsychological literature, serial position data have often been normalized across different lengths, and an improved procedure for this has recently been reported by Machtynger and Shallice (2009). Effects of length and a U-shaped normalized serial position curve have been criteria for identifying working memory deficits. We present simulations and analyses to illustrate some of the issues that arise when relating serial position data to specific theories. We show that critical distinctions are often difficult to make based on normalized data. We suggest that curves for different lengths are best presented in their raw form and that binomial regression can be used to answer specific questions about the effects of length, position, and linear or nonlinear shape that are critical to making theoretical distinctions. © 2010 Psychology Press.

Comparing sentence-level features for authorship analysis in Portuguese

In this paper we compare the robustness of several types of stylistic markers to help discriminate authorship at sentence level. We train a SVM-based classifier using each set of features separately and perform sentence-level authorship analysis over corpus of editorials published in a Portuguese quality newspaper. Results show that features based on POS information, punctuation and word / sentence length contribute to a more robust sentence-level authorship analysis. © Springer-Verlag Berlin Heidelberg 2010.

Intersubjectivity and sentiment:from language to knowledge

Intersubjectivity is an important concept in psychology and sociology. It refers to sharing conceptualizations through social interactions in a community and using such shared conceptualization as a resource to interpret things that happen in everyday life. In this work, we make use of intersubjectivity as the basis to model shared stance and subjectivity for sentiment analysis. We construct an intersubjectivity network which links review writers, terms they used, as well as the polarities of the terms. Based on this network model, we propose a method to learn writer embeddings which are subsequently incorporated into a convolutional neural network for sentiment analysis. Evaluations on the IMDB, Yelp 2013 and Yelp 2014 datasets show that the proposed approach has achieved the state-of-the-art performance.