Bi-view semi-supervised active learning for cross-lingual sentiment classification

Recently, sentiment classification has received considerable attention within the natural language processing research community. However, since most recent works regarding sentiment classification have been done in the English language, there are accordingly not enough sentiment resources in other languages. Manual construction of reliable sentiment resources is a very difficult and time-consuming task. Cross-lingual sentiment classification aims to utilize annotated sentiment resources in one language (typically English) for sentiment classification of text documents in another language. Most existing research works rely on automatic machine translation services to directly project information from one language to another. However, different term distribution between original and translated text documents and translation errors are two main problems faced in the case of using only machine translation. To overcome these problems, we propose a novel learning model based on active learning and semi-supervised co-training to incorporate unlabelled data from the target language into the learning process in a bi-view framework. This model attempts to enrich training data by adding the most confident automatically-labelled examples, as well as a few of the most informative manually-labelled examples from unlabelled data in an iterative process. Further, in this model, we consider the density of unlabelled data so as to select more representative unlabelled examples in order to avoid outlier selection in active learning. The proposed model was applied to book review datasets in three different languages. Experiments showed that our model can effectively improve the cross-lingual sentiment classification performance and reduce labelling efforts in comparison with some baseline methods.     

What improves citizens’ privacy perceptions toward RFID technology? A cross-country investigation using mixed method approach

Privacy is a serious concern to radio frequency identification (RFID) technology. Worldwide, several companies scrapped RFID projects because of high resistance from consumers and their advocacy groups – which actually demand RFID-specific privacy policies. This concern is even more acute when RFID is used in public applications; because, in general case, citizens cannot refuse to provide data, and the data collected by a government agency would offer serious threats if are shared among third parties. Limited research has been performed in this specific issue; they all agree that perceived privacy increased RFID acceptance. But, what drives privacy perceptions are yet to be researched – this study closes this research gap. In order to conduct the current research, mixed method of research approach has been adopted. In the qualitative research stage, the authors conducted two focused-group discussion sessions and eight in-depth interviews in two different countries: Australia and Bangladesh; arguing that the status, and the perceptions and tolerance of the citizens on privacy are different in these two regions. The explored factors have been examined with empirical data obtained from these two countries. It is found that, there are distinct differences in perceptions in developed and developing countries. The detail findings offer practical suggestions to the agency managers so that they can ensure better privacy of the citizens. As a significant theoretical contribution, this study enhances existing literature identifying the antecedents of privacy, which play even different roles in different cultural backgrounds.     

A micropillar array for sample concentration via in-plane evaporation

We present a method to perform sample concentration within a lab-on-a-chip using a microfluidic structure which controls the liquid-gas interface through a micropillar array fabricated in polydimethylsiloxane between microfluidic channels. The microstructure confines the liquid flow and a thermal gradient is used to drive evaporation at the liquid-gas-interface. The evaporation occurs in-plane to the microfluidic device, allowing for precise control of the ambient environment. This method is demonstrated with a sample containing 1 μm, 100 nm fluorescent beads and SYTO-9 labelled Escherichia coli bacteria. Over 100 s, the fluorescent beads and bacteria are concentrated by a factor of 10.     

An empirical investigation of entrepreneurship intensity in Iranian state universities

The purpose of this study is to propose a framework to evaluate the entrepreneurship intensity (EI) of Iranian state universities. In order to determine EI, a hybrid multi-method framework consisting of Delphi, Analytic Network Process (ANP), and VIKOR is proposed. The Delphi method is used to localize and reduce the number of criteria extracted from a deep literature review, according to the social and economic conditions of Iranian state universities by using an expert panel, including sixty-eight country-wide academicians and practitioners. After that, a group approach to ANP was utilized as an evaluation method to derive the weights of each criterion. Next, the evaluation data were gathered through a questionnaire, and, finally, the compromise ranking of universities was calculated using the VIKOR method. Moreover, this study applies weight-variance analysis (WVA) to suggest improvement actions. The paper proposes an evaluation framework for determining the performance of entrepreneurship development initiatives in universities. By using this framework, twenty-two Iranian state universities are evaluated in terms of their EI, and the results are discussed.     

Parents' attitudes towards inclusion of students with autism in Jordan

This study investigated the attitudes of parents in Jordan towards the inclusion of students with autism spectrum disorder (ASD) in public schools and what the parents believed to be the most important prerequisite of child-based skills for successful inclusion. A total of 148 parents were selected to complete the survey. The researchers explored whether variable demographic characteristics such as age, student's gender, parent's gender, education levels, monthly income, and high- or low-function ASD correlated with the attitudes of parents towards inclusion of students with ASD in public schools. The analyses revealed that the variables that correlated with parents' attitudes towards inclusion were education levels and high- or low-function ASD. The themes behind the parents' attitude for inclusion are discussed. With regard to the prerequisite skills for successful inclusion, the results indicated that parents recommended independent skills, playing skills, behavioural skills, imitation skills, routine skills, social skills, paying attention skills, language skills, and pre-academic and academic skills in that order.     

A family of proportionate normalized subband adaptive filter algorithms

In this paper, the concept of proportionate adaptation is extended to the normalized subband adaptive filter (NSAF), and seven proportionate normalized subband adaptive filter algorithms are established. The proposed algorithms are proportionate normalized subband adaptive filter (PNSAF), μ‐law PNSAF (MPNSAF), improved PNSAF (IPNSAF), the improved IPNSAF (IIPNSAF), the set-membership IPNSAF (SM-IPNSAF), the selective partial update IPNSAF (SPU-IPNSAF), and SM-SPU-IPNSAF which are suitable for sparse system identification in network echo cancellation. When the impulse response of the echo path is sparse, the PNSAF has initial faster convergence than NSAF but slows down dramatically after initial convergence. The MPNSAF algorithm has fast convergence speed during the whole adaptation. The IPNSAF algorithm is suitable for both sparse and dispersive impulse responses. The SM-IPNSAF exhibits good performance with significant reduction in the overall computational complexity compared with the ordinary IPNSAF. In SPU-IPNSAF, the filter coefficients are partially updated rather than the entire filter at every adaptation. In SM-SPU-IPNSAF algorithm, the concepts of SM and SPU are combined which leads to a reduction in computational complexity. The simulation results show good performance of the proposed algorithms.     

Optimized frequency-based foundation design for wind turbine towers utilizing soil-structure interaction

This study illustrates design optimization for multiple wind towers located at different villages in Alaska. The towers are supported by two different types of foundations: large mat or deep piles foundations. Initially, a reinforced concrete (RC) mat foundation was proposed. Where soil conditions required it, a pile foundation solution was devised utilizing a 30 in thick RC mat containing an embedded steel grillage of W18 beams and supported by 20-24 in grouted or un-grouted piles. For faster installation and lower construction cost, all-steel foundations were proposed for these remote Alaska sites. The new all-steel design was found to reduce the natural frequencies of the structural system due to softening the foundation. Thus, the tower-foundation system could potentially become near-resonant with the operational frequencies of the wind turbine. Consequently, the likelihood of structural damage or even the collapse is increased.A detailed 3D finite-element model of the tower-foundation-pile system with RC foundation was created using SAP2000. Soil springs were included in the model based on soil properties obtained from the geotechnical investigation. The natural frequency from the model was verified against the tower manufacturer analytical and experimental values. When piles were used, numerous iterations were carried out to eliminate the need for the RC and optimize the design. An optimized design was achieved with enough separation between the natural and operational frequencies. The design successfully avoids damage to the structural system, while eliminating the need for any RC in most cases.     

Enhancing Aspect-Based Sentiment Analysis of Arabic Hotels’ reviews using morphological,syntactic and semantic features

This research presents an enhanced approach for Aspect-Based Sentiment Analysis (ABSA) of Hotels’ Arabic reviews using supervised machine learning. The proposed approach employs a state-of-the-art research of training a set of classifiers with morphological, syntactic, and semantic features to address the research tasks namely: (a) T1:Aspect Category Identification, (b) T2:Opinion Target Expression (OTE) Extraction, and (c) T3: Sentiment Polarity Identification. Employed classifiers include Naïve Bayes, Bayes Networks, Decision Tree, K-Nearest Neighbor (K-NN), and Support-Vector Machine (SVM).The approach was evaluated using a reference dataset based on Semantic Evaluation 2016 workshop (SemEval-2016: Task-5). Results show that the supervised learning approach outperforms related work evaluated using the same dataset. More precisely, evaluation results show that all classifiers in the proposed approach outperform the baseline approach, and the overall enhancement for the best performing classifier (SVM) is around 53% for T1, around 59% for T2, and around 19% in T3.     

An evaluation of an adaptive learning system based on multimodal affect recognition for learners with intellectual disabilities

Artificial intelligence tools for education (AIEd) have been used to automate the provision of learning support to mainstream learners. One of the most innovative approaches in this field is the use of data and machine learning for the detection of a student’s affective state, to move them out of negative states that inhibit learning, into positive states such as engagement. In spite of their obvious potential to provide the personalisation that would give extra support for learners with intellectual disabilities, little work on AIEd systems that utilise affect recognition currently addresses this group. Our system used multimodal sensor data and machine learning to first identify three affective states linked to learning (engagement, frustration, boredom) and second determine the presentation of learning content so that the learner is maintained in an optimal affective state and rate of learning is maximised. To evaluate this adaptive learning system, 67 participants aged between 6 and 18 years acting as their own control took part in a series of sessions using the system. Sessions alternated between using the system with both affect detection and learning achievement to drive the selection of learning content (intervention) and using learning achievement alone (control) to drive the selection of learning content. Lack of boredom was the state with the strongest link to achievement, with both frustration and engagement positively related to achievement. There was significantly more engagement and less boredom in intervention than control sessions, but no significant difference in achievement. These results suggest that engagement does increase when activities are tailored to the personal needs and emotional state of the learner and that the system was promoting affective states that in turn promote learning. However, longer exposure is necessary to determine the effect on learning.