Linguistic Feature Selection Research Articles

A Look at the Nativization of Bangladeshi English through Corpus Data

This paper explores Bangladeshi English, a relative newcomer to the family of world Englishes. First, we chart the evolution of English in Bangladesh across several phases of development. These show that a number of political and ideological factors make the evolution of English in Bangladesh unique in that it wavers between the status of an exonormative foreign language (English in Bangladesh) and the development of a local variety (Bangladeshi English). From a linguistic perspective, recent studies agree that the current level of proficiency is very low, with a dearth of teachers and an absence of quality education. Second, we examine the degree of nativization of Present-day Bangladeshi English on the basis of (i) its postcolonial evolution and the more recent effects of globalization, following the most popular models of analysis (the Dynamic Model and the Extra- and Intra-Territorial Forces Model), and (ii) linguistic evidence obtained through the analysis of a selection of linguistic features associated with this variety, as represented in the GloWbE corpus.

An empiric validation of linguistic features in machine learning models for fake news detection

The diffusion of fake news is a growing problem with a high and negative social impact. There are several approaches to address the detection of fake news. This work focuses on a hybrid approach based on functional linguistic features and machine learning. There are several recent works with this approach. However, there are no clear guidelines on which linguistic features are most appropriate nor how to justify their use. Furthermore, many classification results are modest compared to recent advances in natural language processing. Our proposal considers 88 features organized in surface information, part of speech, discursive characteristics, and readability indices. On a 42 677 news database, we show that the classification results outperform previous work, even outperforming state-of-the-art techniques such as BERT, reaching 99.99% accuracy. A proper selection of linguistic features is crucial for interpretability as well as the performance of the models. In this sense, our proposal contributes to the intentional selection of linguistic features, overcoming current technical issues. We identified 32 features that show differences between the type of news. The results are highly competitive in the classification and simple to implement and interpret.

Priscian’s Latin: new insights from the manuscript tradition

Abstract The paper deals with a selection of linguistic features of Priscian’s Latin (6th century AD), namely the use of the indefinite pronoun and adjective after a negation, the declension of Greek loan-words with nominative ending in -ys, and the use of verbal moods and tenses in concessive clauses. The inquiry takes advantage of the new ongoing collation of the 8th–10th-century manuscripts of the Ars Prisciani, which enables ascription to Priscian’s linguistic use of some peculiar syntactic and morphological features that had been obliterated by the standardizing attitude of the former critical editor, Martin Hertz.

Creativity vs. misinterpretation in translanguaging contexts: perspectives of multilingual/bilingual addressees

ABSTRACT Translanguaging studies have paid much attention to meaning-making processes, exploring multilingual speakers’ strategic selection of linguistic features from a holistic linguistic repertoire to convey meanings, and assuming that multilingual addressees can successfully decode the encoded meanings. Failure in the meaning-interpreting processes in relation to translanguaging contexts is largely ignored. To bridge the gap, this study explores Chinese multilingual/bilingual addressees’ misinterpretations in translanguaging contexts, by probing into the mechanisms of misinterpretation. Misinterpretations are found to occur as a consequence of gaps between the addresser and the addressee in the cognitive, social-historical, and cultural dimensions of a Translanguaging Space, and failure of the Translanguaging Instinct resulting from various causes, such as confusion of and preferences for different linguistic codes. Furthermore, Chinese addressees’ misinterpretations are also related to addressers’ creativity or novel expressions and the spreading use of pinyin/English abbreviations in online communication. With this study, it is argued that translanguaging as a theory of practice should look into how the meaning of a new word, phrase, or abbreviation becomes conventionalised and how newcomers of a community are socialised into the shared practice of translanguaging.

HSI-LFS-BERT: Novel Hybrid Swarm Intelligence Based Linguistics Feature Selection and Computational Intelligent Model for Alzheimer’s Prediction Using Audio Transcript

Alzheimer’s dementia (AD) affects memory, language, and cognition which worsens over time. It is critical to develop a reliable early detection method before permanent brain atrophy and cognitive impairment. This study uses clinical transcripts, a text-based adaptation of the original audio recordings of Alzheimer’s patients. This audio transcript data is taken from DementiaBank which is the largest public dataset of AD transcripts. The study aims to show how transfer learning-based models and swarm intelligence optimization techniques can be used to predict Alzheimer’s disease. To enhance the prediction performance for Alzheimer’s disease, a hybrid swarm intelligence linguistic feature selection (HSI-LFS) approach is proposed that extracts a combined feature set using Particle Swarm Optimization (PSO), Dragonfly Optimization (DO), and Grey Wolf Optimization (GWO) algorithms. In addition, a transfer learning-based model called HSI-LFS-BERT, a combination of the HSI-LFS feature selection method and Bidirectional Encoder Representations from Transformer (BERT) algorithm is proposed. The proposed model is compared using two feature sets: first set consists of the initial feature set and the second set contains hybrid feature set that has been extracted using the suggested HSI-LFS method. The BERT embedding with HSI-LFS outperformed the conventional feature set providing the most accurate modeling parameters while reducing the computations by 27.19%. The proposed HSI-LFS-BERT model excelled state-of-the-art models achieving 98.24% accuracy, 91.56% precision, and 98.78% recall.

Chineseness and Cantonese tones in post-1997 Hong Kong

This article explores the ideological selection of linguistic features as emblems. Building on the notion of raciolinguistic enregisterment, it examines why the Cantonese tone system, rather than some other linguistic feature, has lately transformed into an emblem of the Han race, Chineseness, and Hongkongers. Three characteristics of Cantonese tones facilitate this transformation: their multifaceted nature, their indexical expansiveness, and the ‘storiability’ of their indexical meanings. The indexical linkages that Cantonese tones evoke fit into a myth of Cantonese superiority, which speaks to many Hongkongers' anxiety resulting from the recent rise of mainland China and the concomitant decline of their city.

Systematicity in linguistic feature selection: Repair sequences and subsequent accommodation

AbstractThis study examines linguistic feature selection and its relationship with repair sequences in a longitudinal corpus of Japanese–Filipino business ELF interactions. In the corpus, Japanese employees communicate once a month with Filipino employees via computer software to confirm infrastructure status at a Filipino company’s factories. Comparative constructions frequently appear in the corpus because of the nature of the interactions, but the kinds and frequencies of comparative constructions change month to month. This study demonstrates that early in the corpus, the speakers utilized a multitude of comparative constructions, but after 12 months, the speakers have settled on one preferred comparative construction. Furthermore, the preferred construction emerged from repair sequences, which suggests that repair is significantly related to linguistic feature selection. Accordingly, this study hypothesizes that repair sequences do far more than just resolve an interactional problem; repaired linguistic features are more likely to be selected again the next time a similar linguistic feature is relevant to the progression of the interaction.

Linguistic analysis of datasets for semantic textual similarity

Abstract Semantic Textual Similarity (STS), which measures the equivalence of meanings between two textual segments, is an important and useful task in Natural Language Processing. In this article, we have analyzed the datasets provided by the Semantic Evaluation (SemEval) 2012–2014 campaigns for this task in order to find out appropriate linguistic features for each dataset, taking into account the influence that linguistic features at different levels (e.g. syntactic constituents and lexical semantics) might have on the sentence similarity. Results indicate that a linguistic feature may have a different effect on different corpus due to the great difference in sentence structure and vocabulary between datasets. Thus, we conclude that the selection of linguistic features according to the genre of the text might be a good strategy for obtaining better results in the STS task. This analysis could be a useful reference for measuring system building and linguistic feature tuning.

Data Classification by using Homogeneous Clustering Process

In this paper we have performed dynamic clustering based on classification. The Enhance Neuro-fuzzy system for classification using dynamic clustering presented in this paper is an extension of the original Neuro-fuzzy method for linguistic feature selection and rule-based classification. In the dynamic clustering process, the parameter values like centroid, threshold value and standard deviation are estimated. This parameter is used for creating the cluster. The Gaussian membership function is applied to these clusters to generate the binary value of each feature to given cluster. Using this method we have got the large number of cluster and minimum accuracy. To reduce the cluster size and to improve the accuracy we have implement the homogenous clustering process. Using this process we have minimize the cluster size, and also improve the accuracy. General Terms Fuzzy logic, Artificial intelligence, Neural Network.

EU discourse: Polyphony and unclearness

The present article discusses the relevance of studying EU political discourse in a linguistically based polyphonic perspective. The claim is that polyphony is a fruitful approach in the sense that it reveals different types of hidden interaction and thus uncovers implicit and unclear messages. Examples are taken from two speeches given by former Prime Minister Tony Blair, on Europe and European integration, speeches which can be characterised as ‘visionary speeches’. A selection of linguistic features indicating the presence of different voices in these speeches, which are formally monological, will be presented. The analysis will thus help to identify the nature of the postulated unclearness in political discourse and to make explicit the complex relationship between text and context. The theoretical framework used for analysing the speeches will be linguistic polyphony as developed in the ScaPoLine theory.

A Hierarchical Neural Network Document Classifier with Linguistic Feature Selection

In this article, a neural network document classifier with linguistic feature selection and multi-category output is presented. It consists of a feature selection unit and a hierarchical neural net...

Examining the “voice of interpreting” in speech pathology

This paper investigates professional interpreting practice in the setting of speech pathology through a multifaceted analysis of the transcripts of three recorded sessions involving first-generation Italian-speaking immigrants to Australia and English-speaking healthcare professionals working in Melbourne. Applying Mishler’s notion of “voice” to the context of interpreter-mediated communication and focusing on a selection of linguistic features — ranging from turn-taking and topic development to the interpreter’s choice of footing, departures from the primary speakers’ utterances, and use of prosodic resources — the discussion identifies the voice that interpreters, as third participants in the interaction, choose to adopt between the “voice of medicine” and the “voice of the lifeworld”. The study is of a qualitative nature, although a general indication of the frequency of certain features is supplied, and interpreting conduct is described rather than prescribed. The reporting and interpretation of findings are, however, informed by and reflect issues of value revolving around the concept of “humane medical care”.

CONVERGENCE AND THE FORMATION OF AFRIKAANS

As a phenomenon to be explained, convergence in historical linguistics is substantively no different than in creolistics. The general idea is that accommodation by speakers of “established” languages in contact and the formation of new language varieties both involve a process of leveling of different structures that achieve the same referential and nonreferential effects. The relatively short and well-documented history of Afrikaans presents an important case study in the competition and selection of linguistic features during intensive language contact.