Word Clusters Research Articles

Effects of metrical dissonance and expertise on perceived emotion in Schumann’s Carnaval

Numerous studies have investigated the effects of pitch structures on perceived emotion in music, but the emotional effects of rhythm and meter have received far less attention. In the experiment reported here, we manipulated the metrical framework of music by Robert Schumann and asked participants to judge perceived emotion in the resulting excerpts. The distinction between metrical dissonance and consonance offered by Harald Krebs (1999) was the theoretical basis of this study. Stimuli were 10 metrically dissonant excerpts from Schumann’s Carnaval and 10 metrically consonant recompositions of these excerpts. Recompositions maintained original tempi, global meters, and harmonic frameworks. Participants—graduate-level pianists and non-musicians—heard all 20 excerpts in randomized order. On each trial, they chose a cluster of emotion words, based on Schubert (2003), that best described the excerpt, a cluster that worst described the excerpt, and rated their level of interest. Results indicate significant effects of both metrical character and expertise on perceived emotion. There were also significant differences in the likelihood of participants changing their responses between metrically dissonant and consonant versions as a function of musical excerpt. This last finding leads to suggestions for future investigations on degrees of metrical dissonance.

Geo-spatial text-mining from Twitter – a feature space analysis with a view toward building classification in urban regions

ABSTRACTBy the year 2050, it is expected that about 68% of global population will live in cities. To understand the emerging changes in urban structures, new data sources like social media must be taken into account. In this work, we conduct a feature space analysis of geo-tagged Twitter text messages from the Los Angeles area and a geo-spatial text mining approach to classify buildings types into commercial and residential. To create the feature space, broadly accepted word embedding models like word2vec, fastText and GloVe as well as more traditional models based on TF-IDF have been considered. A visual analysis of the word embeddings shows that the two examined classes yield several word clusters. However, the classification results produced by Naïve Bayes support vector machines, and a convolutional neural network indicates that building classification from pure social media text is quite challenging. Furthermore, this work illustrates a base toward fusing text features and remote sensing images to classify urban building types.

The advantages and challenges of “big data”: Insights from the 14 billion word iWeb corpus

The iWeb corpus contains nearly 14 billion words from 22 million web pages, and it has been designed in a way that allows users to quickly and easily create “Virtual Corpora”, in order to focus on websites that are related to their areas of interest. The data from this very large corpus provides very detailed information on syntactic, morphological, lexical, and semantic phenomena, in ways that would never be possible with a small 100 million or 500 million word corpus. In addition, the corpus provides a number of features that are not available with other large corpora, such as the ability to perform advanced searches of the top 60,000 words in the corpus, and to see a wealth of information on each of these words - definitions, links to images and audio, translations, detailed frequency information, related topics, collocates, word clusters, re-sortable concordance lines, and much more. Finally, we discuss the challenges of large corpora, and how the corpus architecture that is used for iWeb has uniquely been designed to address these challenges. (Brigham Young University · Kyung Hee University)

Recognition and extraction of named entities in online medical diagnosis data based on a deep neural network

How to correctly recognize and extract named entities, such as disease names, medical measurements and therapies, from online medical diagnosis data remains challenging. For the one hand, conventional natural language processing methods cannot be directly applied in the field of online medical diagnosis (OMD). Although existing supervised or unsupervised learning algorithms have offered strategies for OMD on open websites, such methods might extensively rely on specific knowledge sources or manually designed features. For the other hand, due to the large size of the data and the sophistication of the data structure, it is difficult to establish a robust NLP model for recognition and extraction of clinical named entities in paragraphs. Therefore, in the paper, we try to establish a new deep neural network (DNN), combine the bi-directional long short-term memory (Bi-LSTM) and the conditional random field (CRF), and utilize online medical diagnosis data for recognition and extraction of clinical named entities. Different from existing artificial neural networks (ANN), the proposed neural network well functions even without manual rules or features. The word representations input into the DNN is connected by character-based representations and continuous bags of word clusters (CBOWC) in an embedding way. We test the new DNN on different online medical diagnosis datasets obtained from a scalable web crawler, and compare it with the long short-term memory (LSTM) neural network linguistic model, the convolutional neural network (CNN) model and the most advanced Bi-LSTM-CRF. The fundamentals for the method selection is that they have been suggested to generate acceptable results. According to the comparison analyses, the DNN is proved to be a reliable tool and improves every benchmark performance of OMD.

ENHANCEMENT OF TEXT BASED EMOTION RECOGNITION PERFORMANCES USING WORD CLUSTERS

Human Computer Interaction (HCI) researches the use of computer technology mainly focused on the interfaces between human users and computers. Expression of emotion comprises of challenging style as it is produced with plaint text and short messaging language as well. This research paper investigates on the overview of emotion recognition from various texts and expresses the emotion detection methodologies applying Machine Learning Approach (MLA). This paper recommends resolving the problem of feature meagerness, and largely improving the emotion recognition presentation from short texts by achieving the three aims: (I) The representing short texts along with word cluster features, (II) Presenting a narrative word clustering algorithm, and (iii) Making use of a new feature weighting scheme of the Emotion classification. Experiments were performed for the classifying the emotions with different features and weighting schemes, on the openly available dataset. We have used the word clusters in place of unigrams as features, the micro-averages of accuracy have been found to be enhanced by more than three percentage, which suggests that the overall accuracy value of the text emotion classifier has been improved. All the macro-averages were enhanced by more than one percentage, which suggests that the word cluster feature can advance the generalization potential of the emotion classifier. The experimental results suggest that the text words cluster features and the proposed weighting scheme can moderately resolve the problems of the emotion recognition performance and the feature sparseness.

Indexing the Theologico-Political

Political theology is an important and growing field of theological research, which has also attracted the attention of political scientists, philosophers, and social theorists. This interest has led to different and often ambiguous definitions of what political theology actually is. In addition, related subfields like liberation theology, public theology, and critical theory are often associated with political theology, making categorization challenging. It is important that theological librarians understand the changing terrain of theological-political inquiry in order to better provide access for their research communities.
 In this paper, we present findings about the range of current description practices for books and journal articles related to political theology. We share preliminary explorations in mining cataloged resources for subject headings related to political theology, as well as full text of titles in HathiTrust for clusters of words that can be modeled as topics. These tools can help librarians to better understand the research literature they are describing, and engage with it more actively.

Softly Associative Transfer Learning for Cross-Domain Classification.

The main challenge of cross-domain text classification is to train a classifier in a source domain while applying it to a different target domain. Many transfer learning-based algorithms, for example, dual transfer learning, triplex transfer learning, etc., have been proposed for cross-domain classification, by detecting a shared low-dimensional feature representation for both source and target domains. These methods, however, often assume that the word clusters matrix or the clusters association matrix as knowledge transferring bridges are exactly the same across different domains, which is actually unrealistic in real-world applications and, therefore, could degrade classification performance. In light of this, in this paper, we propose a softly associative transfer learning algorithm for cross-domain text classification. Specifically, we integrate two non-negative matrix tri-factorizations into a joint optimization framework, with approximate constraints on both word clusters matrices and clusters association matrices so as to allow proper diversity in knowledge transfer, and with another approximate constraint on class labels in source domains in order to handle noisy labels. An iterative algorithm is then proposed to solve the above problem, with its convergence verified theoretically and empirically. Extensive experimental results on various text datasets demonstrate the effectiveness of our algorithm, even with the presence of abundant state-of-the-art competitors.

Research on Domain Term Dictionary Construction Based on Chinese Wikipedia

Domain terms are words or phrases that represent concepts or relationships in a specific domain. It can represent the characteristics of corresponding domains. The automatic construction of domain-specific dictionary is an important task in natural language processing, which can be adopted in domain-specific ontology construction, vertical search, text classification, information retrieval, question answering system etc. In this paper, we propose a novel method for constructing domain term dictionary based on Chinese Wikipedia web resource and deep learning technology. We for first time explore to word representation by Word2vec model integrating Wikipedia link structure. Then we use word clustering algorithm and seed word extraction method to construct an original domain dictionary. Moreover, neural network method is applied to extend domain dictionary. In experiments, different methods were employed to extract the domain-specific terms and their performances were compared in automobile field, the results reveal the effectiveness of our method for construction of domain-specific dictionary.

Research on Hot Topics of College English Education in China

In order to study the hot topics of the college English and lay the foundation for the next research, the article based on 2017 published journal in CNKI (Chinese National Knowledge Infrastructure) on the English study in our university academic thesis co word clustering. After the preliminary analysis, the research focuses on seven aspects: the countermeasure analysis, the multimedia teaching, the network teaching, the cross culture teaching, the curriculum, the quality education and the teaching reform. Therefore, to promote the development of College English teaching, we must clear to expand the trend of hot field, reduce overheating research field, transfer the research direction to other neglected field in various fields so that we can promote the coordinated development.

Document representation based on probabilistic word clustering in customer-voice classification

Customer-voice data have an important role in different fields including marketing, product planning, and quality assurance. However, owing to the manual processes involved, there are problems associated with the classification of customer-voice data. This study focuses on building automatic classifiers for customer-voice data with newly proposed document representation methods based on neural-embedding and probabilistic word-clustering approaches. Semantically similar terms are classified into a common cluster. The words generated from neural embedding are clustered according to the membership strength of each word relative to each cluster derived from a probabilistic clustering method such as the fuzzy C-means clustering method or Gaussian mixture model. It is expected that the proposed method can be suitable for the classification of customer-voice data consisting of unstructured text by considering the membership strength. The results demonstrate that the proposed method achieved an accuracy of 89.24% with respect to representational effectiveness and an accuracy of 87.76% with respect to the classification performance of customer-voice data consisting of 12 classes. Further, the method provided an intuitive interpretation for the generated representation.

A Systematic Mapping Study on Customer Loyalty and Brand Management

Customer loyalty is a topic of great interest for marketing scholars due to its importance in gaining sustainable competitive advantages and financial outcomes. Literature is prolific of works regarding customer loyalty and brand management. In order to improve the quantity and quality of research publications, research themes mapping of published studies is required. The aim of this paper is to provide scholars with a systematization and mapping of the contributions on this topic to develop an effective research road map for future research. A bibliometric analysis and a mapping study have been conducted on 337 publications on customer loyalty and brand management from 2000 to 2018. The results present the most cited works on the topic, an evaluation map showing the most frequent and cited words and six clusters of words based on their co-occurrence. From the analysis of the clusters, the most relevant research findings, trends, and issues emerge, suggesting interesting theoretical and practical implications.

Social Media Text Data Visualization Modeling: A Timely Topic Score Technique

Due to the rapid growth of large size text data from Internet sources like Twitter, social media platforms have become the more popular sources to be utilized to extract information. The extracted text information is then further converted to number through a series of data transformation and then analyzed through text analytics models for decision-making problems. Among the text analytics models, one particular common and popular one is based on Latent Dirichlet Allocation (LDA), which is a topic model method with the topics being clusters of words in the documents associated with fitted multivariate statistical distributions. However, these models are often poor estimators of topic proportions. Hence, this paper proposes a timely topic score technique for social media text data visualization, which is based on a point system from topic models to support text signaling. This importance score system is intended to mitigate the weakness of topic models by employing the topic proportion outputs and assigning importance points to present text topic trends. The technique then generates visualization tools to show topic trends over the studied time period and then further facilitate decision-making problems. Finally, this paper studies two real-life case examples from Twitter text sources and illustrates the efficiency of the methodology.

Analysis of Experiments on Statistical and Neural Parsing for a Morphologically Rich and Free Word Order Language Urdu

This article presents an analysis of experiments with statistical and neural parsing techniques for Urdu, a widely spoken South Asian language. We demonstrate state of the art constituency parsing results for an Urdu treebank. Urdu is a morphologically rich and is characterized by free word order. Language representation (e.g. input type, lemmatization, word clusters), part of speech tag set, phrase labels and the size of a training corpus are crucial for parsing such languages. In this article, probabilistic context-free grammars, data-oriented parsing, and recursive neural network based models have been experimented with several linguistic features which show improvements in the parsing results. Features include syntactic sub-categorization of POS tags, empirically learned horizontal and vertical markovizations and lexical head words. These features enable dependency information for case markers and add phrasal and lexical context to the parse trees. The data-oriented parsing and recursive neural network model give an f-score of 87.1 by considering gold POS tags in the test set, on textual input, they show a performance with f-scores of 83.4 and 84.2, respectively. To overcome the issue of data sparsity due to the morphological richness, lemmatization and unsupervised word clustering have been performed. A treebank should cover most probable word orders of the language so that models can learn various orders accurately. To analyze the order coverage of the treebank and learning capability of different parsers, a test set has been prepared conditioning different word orders. This test set is evaluated with the best performing parsing models and with gold POS tags, f-scores are above 90 and on textual input, the average f-score is 87.6.

Sonority Sequencing Principle in Sabzevari Persian: A Constraint-Based Approach

Abstract This study sheds light on the relationship between the Sonority Sequencing Principle (SSP) and syllable structure in Sabzevari, a Persian vernacular spoken in the Sabzevar area of Northeast Iran. Optimality Theory (OT), as a constraint-based approach, is utilized to scrutinize sonority violation and its repair strategies. The results suggest that obedience to the SSP is mandatory in Sabzevari, as shown through the treatment of word-final clusters in Standard Persian words which violate the SSP. These consonant clusters are avoided in Sabzevari by two phonological processes: vowel epenthesis and metathesis. Vowel epenthesis is motivated by final consonant clusters of the forms /fricative+coronal nasal/, /plosive+bilabial nasal/, /fricative+bilabial nasal/, /plosive+rhotic/, /fricative+rhotic/, and /plosive+lateral/. Metathesis, as another repair strategy for sonority sequencing violations, occurs when dealing with final consonant clusters of the forms /plosive+fricative/and / fricative+lateral/.

Two-Stage Topic Extraction Model for Bibliometric Data Analysis Based on Word Embeddings and Clustering

Topic extraction is an essential task in bibliometric data analysis, data mining and knowledge discovery, which seeks to identify significant topics from text collections. The conventional topic extraction schemes require human intervention and involve also comprehensive pre-processing tasks to represent text collections in an appropriate way. In this paper, we present a two-stage framework for topic extraction from scientific literature. The presented scheme employs a two-staged procedure, where word embedding schemes have been utilized in conjunction with cluster analysis. To extract significant topics from text collections, we propose an improved word embedding scheme, which incorporates word vectors obtained by word2vec, POS2vec, word-position2vec and LDA2vec schemes. In the clustering phase, an improved clustering ensemble framework, which incorporates conventional clustering methods (i.e., k-means, k-modes, k-means++, self-organizing maps and DIANA algorithm) by means of the iterative voting consensus, has been presented. In the empirical analysis, we analyze a corpus containing 160,424 abstracts of articles from various disciplines, including agricultural engineering, economics, engineering and computer science. In the experimental analysis, performance of the proposed scheme has been compared to conventional baseline clustering methods (such as, k-means, k-modes, and k-means++), LDA-based topic modelling and conventional word embedding schemes. The empirical analysis reveals that ensemble word embedding scheme yields better predictive performance compared to the baseline word vectors for topic extraction. Ensemble clustering framework outperforms the baseline clustering methods. The results obtained by the proposed framework show an improvement in Jaccard coefficient, Folkes & Mallows measure and F1 score.

Everyday Words and the Character of Prose in Nineteenth-Century Britain by Jonathan Farina

Reviewed by: Everyday Words and the Character of Prose in Nineteenth-Century Britain by Jonathan Farina Michael Hollington Jonathan Farina. Everyday Words and the Character of Prose in Nineteenth-Century Britain. Cambridge: Cambridge UP, 2017. Pp. xxii + 286. £75. ISBN 9781107181632 Alighting on the phrase "a curate's egg" to characterize my overall reaction to this book, I decided (in the hope of gaining precision in the use of it) to satisfy a longstanding curiosity about its history, and so went online. What I found there was instructive: the phrase originates in an 1895 George du Maurier cartoon in Punch, which depicts a timid curate at table with a bishop. The latter regrets that the curate seems to have been given a bad egg. Anxious to please, the curate replies "'Oh no, my Lord, I assure you! Parts of it are excellent!'" From this, I learnt that the sense in which I want to use the phrase here (like that in which it is mostly used nowadays) is not that of the original joke. Parts of Farina's book are indeed excellent–unlike the curate's egg–but I found it less than thoroughly satisfactory as a whole. Dickens is a constant presence, as one would expect of a book devoted to "everyday words" in the nineteenth century, but to my mind the best chapter in the book is principally about Jane Austen. It pays attention to a cluster of words in the novels–"attentions," "particular" and "general"–from an epistemological point of view, one that distinguishes Farina's perspective throughout the book. Providing us with a valuable sketch of the progress of pragmatism in the eighteenth century (constituting a worthwhile contribution to the intellectual history of that period in relation to the rise of the novel), Farina offers acute commentary on Jane Austen's fine calibrations of the "particular" and the "general," going beyond Byron's Alexander "Suwarrow, who but saw things in the gross–/ Being much too gross to see them in detail" (Don Juan VII, 77) to achieve a razor edge of balance between these binaries. He convincingly proposes this as one source of her universal appeal as a novelist. Everything is right in this chapter–the choice of relatively unambiguous "everyday words" (at least, through their pairing in opposition), the firm grounding in scholarship, the close reading of the relevant texts. Alas, this [End Page 185] is not everywhere the case. The very first chapter, for instance, entitled "Darwin's View from Todgers's: 'A decided turn' for Character and Common Words," fails to convince. It seeks to establish a connection between the repeated use of the word "turn" in the enumeration of Todgers's lodgers–"a gentleman of a sporting turn," "of a debating turn," or of "a smoking turn," etc.–and Darwin's overtly quite different use of the word in expressions like "may well turn the nicely-balanced scale in the struggle of life," or "let us now turn to the nectar-feeding insects"–and, I believe, fails. Although Farina partly redeems himself at the epistemological level again–the comparison between Darwin and Dickens as classifiers and taxonomists of different species within a genus is apposite, for example–there is much vagueness and imprecision in this chapter. Paying attention to "everyday words" seems at times to imply a rejection of analytic close reading, and, concomitantly, something like tone deafness to Dickensian humor. The comparison he makes between the Todgers bachelors and Odysseus the "man of many turns" can surely only make sense as mock-epic, but he solemnly proclaims that they achieve the same "versatility and variety collectively." Nowhere does he seem to investigate the humorous primary meaning and connotation of Dickens's "turn" here–that these are wannabes who dabble in separate favorite hobby-horses (thoroughly traditional in its descent from Sterne etc.) and are very remote from Darwin in every respect except the purely formal resemblance between collections of species. A much better chapter on Dickens, belonging for the most part amongst the good parts of the egg, concerns the conditional analogy "as if." I have long looked forward to reading an in-depth discussion of this quintessential Dickensian phrase, and...

Big Data Analysis in Social Networks for Managing Risks in Clothing Industry

This paper demonstrates the power of big data analysis of social media for the identification and prioritization of supply chain risks in clothing industry. Due to their high volume, data from social media can be considered a good source for obtaining customer feedback. On the other hand, the high volume of social media data could be efficiently processed solely by special tools for big data analysis. Results of such an analysis include clusters of words, topics and sentiments which allow decision makers to gather negative customer feedback and to manage risks in the distribution of products. A case study in the segment of footwear supply chain was analyzed, where one month of data from Twitter was used.

Detecting Adverse Drug Events with Rapidly Trained Classification Models.

IntroductionIdentifying occurrences of medication side effects and adverse drug events (ADEs) is an important and challenging task because they are frequently only mentioned in clinical narrative and are not formally reported.MethodsWe developed a natural language processing (NLP) system that aims to identify mentions of symptoms and drugs in clinical notes and label the relationship between the mentions as indications or ADEs. The system leverages an existing word embeddings model with induced word clusters for dimensionality reduction. It employs a conditional random field (CRF) model for named entity recognition (NER) and a random forest model for relation extraction (RE).ResultsFinal performance of each model was evaluated separately and then combined on a manually annotated evaluation set. The micro-averaged F1 score was 80.9% for NER, 88.1% for RE, and 61.2% for the integrated systems. Outputs from our systems were submitted to the NLP Challenges for Detecting Medication and Adverse Drug Events from Electronic Health Records (MADE 1.0) competition (Yu et al. in http://bio-nlp.org/index.php/projects/39-nlp-challenges, 2018). System performance was evaluated in three tasks (NER, RE, and complete system) with multiple teams submitting output from their systems for each task. Our RE system placed first in Task 2 of the challenge and our integrated system achieved third place in Task 3.ConclusionAdding to the growing number of publications that utilize NLP to detect occurrences of ADEs, our study illustrates the benefits of employing innovative feature engineering.

Global trends of coagulation for water and wastewater treatment by utilizing bibliometrics analysis

Coagulation is one of the most extensively used techniques for drinking water, domestic wastewater, and industrial wastewater treatment due to its low cost and easy operation. The existing strict requirements on water quality facilitated the development of coagulation techniques and researches. Moreover, design and preparation of novel efficient coagulants have been emerging as a significant research topic in the past decades. This study quantitatively assessed current research trends on coagulation for water and wastewater treatment from 1992 to 2016 according to the Science Citation Index-Expanded (SCI-Expanded). Studies related to coagulation for water and wastewater treatment have been significantly increased from 1992 to 2016. Meanwhile, the word cluster analysis is utilized to trace the hotspot related to coagulation for water and wastewater treatment. The investigation showed two main conclusions. One is that the key pollutants in drinking water and wastewater are organic matter, suspensions, and humic substances, especially natural organic matter. The traditional coagulation technique was unsuitable with the continual increase of natural organic matter concentration and the changes of composition. Therefore, it is necessary to combine the conventional coagulation technique with the new technology to adapt to the complex situation of water, which is also the second conclusion. And the new technologies that can be combined with coagulation technique include membranes (nanofiltration, microfiltration, ultrafiltration) and electrocoagulation. The integration of traditional technology and new technology is also the trendency of water treatment in the future.

Urban governance in Latin America: Bibliometrics applied to the context of smart cities

Abstract Society lives in a dynamic, complex, and contradictory urban system that seeks the balance between urban development and environmental preservation. For this purpose, there are Information and Communication Technologies which can make the existence of Smart Cities possible. Thus, developing Smart Cities involves applying inter and multidisciplinary knowledge, which permeates even Information Science. This study analyzes Latin American scientific production by applying Bibliometrics to quantify the publications related to smart cities, geotechnologies, governance, and cadastre. We analyzed scientifi c productionindexed in the Scopus database from 2007 to 2017. We also performed the Factorial Correspondence Analysis for visualization of word clusters in order to verify their co-occurrence in the abstracts of articles. The results identifi ed Brazil as the country with the highest number of published documents. The importance of the term “Smart Cities” was evidenced with the increase of production in the last 5 years, a period that accumulates 98% of such publications. The most common term is “Geographic Information System”, that appears in 75% of the publications. Finally, in the co-occurrence of words, we identifi ed terms that corroborate with the greater objectives of smart cities.