Word Position Research Articles

Aspect-Level Sentiment Analysis Using CNN Over BERT-GCN

The increase in the volume of user-generated content on Twitter has resulted in tweet sentiment analysis becoming an essential tool for the extraction of information about Twitter users’ emotional state. Consequently, there has been a rapid growth of tweet sentiment analysis in the area of natural language processing. Tweet sentiment analysis is increasingly applied in many areas, such as decision support systems and recommendation systems. Therefore, improving the accuracy of tweet sentiment analysis has become practical and an area of interest for many researchers. Many approaches have tried to improve the performance of tweet sentiment analysis methods by using the feature ensemble method. However, most of the previous methods attempted to model the syntactic information of words without considering the sentiment context of these words. Besides, the positioning of words and the impact of phrases containing fuzzy sentiment have not been mentioned in many studies. This study proposed a new approach based on a feature ensemble model related to tweets containing fuzzy sentiment by taking into account elements such as lexical, word-type, semantic, position, and sentiment polarity of words. The proposed method has been experimented on with real data, and the result proves effective in improving the performance of tweet sentiment analysis in terms of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">F</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> score.

Монгол хэлний дуугүй хамжих гийгүүлэгчийг туршин шинжилсэн дүн

In this article, the author analyzes the acoustic-phonetic features of the consonants /b/, /d/, and /g/ of the Khalkha dialect, which is the standard dialect of the Mongolian Literary Language. On this basis, the author explains as follows: The phoneme /b/ is pronounced as the voiceless bilabial plosive [p] (as in stop [paɬ] ‘pen’) in any word position. The phoneme /d/is pronounced as the voiceless dental plosive [t] (as in stop [tawaː] ‘mountain pass’) in any word position. The voiceless velar plosive [k] (as in stop [kerəɬ] ‘light’) and the voiceless uvular plosive [q] (as in stop [qaɬoː] ‘бал’) are allophones for the phoneme /g/ as the initial consonant of a word or after [ŋ] at the beginning of noninitial syllables.

О методе распознавания голосовых команд с применением особого преобразования спектральных плотностей

This article discusses new approach to solving specific problem of speech recognition. The problem is formulated as problem of command recognition. Commands have equal lengths (in words). Each word position has its own set of word candidates. There were developed the models for solving this problem. The models consist of spectral density estimator, convolutional neural networks and naive Bayes classifier. The author performed a comparative analysis of the models and selected the best. There was developed a graphical user interface for interacting with the best model. The result can be used as a base for creation speech recognition methods for voice user interfaces.

Интонационное оформление диктальных вопросов со словом кем ‘кто’ в алтайском языке

This paper presents the experimental phonetic analysis of the Altai language intonation in terms of a communicative focus of the utterance. The interrogative intonation implementation was studied in sounding texts recorded from six informants, all representatives of the Altai-Kizhi dialect. In Altai, as in other Turkic languages, interrogative utterances are mainly expressed in lexical-grammatical and intonational ways. The first involves using interrogative pronouns such as kem “who,” ne “what,” kandyy “what, which,” kancha “how much,” and others, and the interrogative particle ba. The phonetic analysis revealed a wide intonational variability of dictum interrogative utterances with the interrogative pronoun kem. The intonation was found to depend on the interrogative word position. With kem at the beginning, the utterance has an even tone or is marked by an ascending-descending change in the frequency of the basic tone. With kem in the end, the ascending movement of the tone is observed, suggesting the utterance incompleteness and the expectation of the listener’s response. With kem in the middle position, a descending tone is observed, the intonational contour being wavelike: ascending-descending-ascending-descending. Overall, the intonational contour of questions with the word kem can be characterized as descending-rising, but with kem at the very end, the tone is ascending, the frequency difference of the basic tone in all the examples ranging from 5 to 7 pt. The next step will be to specify the trends identified by expanding the research corpus and describing the general pattern of intonation of dictum questions in statements with different interrogative words.

Acoustic classification of coronal stops of Eastern Punjabi.

Punjabi is an Indo-Aryan language which contrasts a rich set of coronal stops at dental and retroflex places of articulation across three laryngeal configurations. Moreover, all these stops occur contrastively in various positions (word-initially, -medially, and -finally). The goal of this study is to investigate how various coronal place and laryngeal contrasts are distinguished acoustically both within and across word positions. A number of temporal and spectral correlates were examined in data from 13 speakers of Eastern Punjabi: Voice Onset Time, release and closure durations, fundamental frequency, F1-F3 formants, spectral center of gravity and standard deviation, H1*-H2*, and cepstral peak prominence. The findings indicated that higher formants and spectral measures were most important for the classification of place contrasts across word positions, whereas laryngeal contrasts were reliably distinguished by durational and voice quality measures. Word-medially and -finally, F2 and F3 of the preceding vowels played a key role in distinguishing the dental and retroflex stops, while spectral noise measures were more important word-initially. The findings of this study contribute to a better understanding of factors involved in the maintenance of typologically rare and phonetically complex sets of place and laryngeal contrasts in the coronal stops of Indo-Aryan languages.

Positional asymmetries in consonant production and intelligibility in dysarthric speech

ABSTRACT Consonant production errors are common in dysarthric speech, but not all consonants are affected to the same extent. Currently, only limited knowledge exists regarding whether different positional allophones are affected to varying degrees in dysarthric speech. The aim of this study was to investigate the effect of consonants’ position-in-words on consonant production accuracy and their relevance to speech intelligibility. To this end, the percentage of correctly articulated consonants was analyzed with respect to position-in-words, manner of articulation, and speakers’ overall intelligibility in dysarthric speech secondary to cerebral palsy. Results revealed that consonants were generally more accurate in initial positions than other positions, and when they were singletons rather than clusters. However, fricatives, a category commonly noted for frequent misarticulations, exhibited no significant positional effects, indicating that fricatives were affected to a similar degree across all word positions. In addition, positional asymmetry manifested to a greater degree as intelligibility decreased. Finally, the strength of the relationship between consonant production accuracy and intelligibility varied depending on positions-in-words, with strong correlations in the medial and final cluster positions but no significant correlation for fricatives in the initial and final singleton positions. The correlation was markedly low for the initial consonants, possibly due to the resistance of initial consonants to production errors. The positional asymmetry found in this study necessitates more research on non-initial consonants and clusters for their potential in improving consonant production and speech intelligibility as a part of dysarthria management.

Effect of linguistic factors on the occurrence of stuttering-like disfluency among Japanese-speaking preschool children who stutter

ABSTRACT This study aimed to investigate the linguistic factors involved in stuttering among Japanese-speaking preschool children. The participants included 10 Japanese children who stutter, with a mean age of 5 years and 9 months. Speech samples comprised spontaneous conversations of the participants with their parents for about 20 minutes. We compared the percentages of the occurrence of stuttering-like disfluencies (SLDs) at the word and sentence levels, using the Wilcoxon signed-rank test. The results showed no significant differences in SLDs based on syllable structure when comparing light and heavy syllables and comparing consonants and vowels in the initial position of each content word. SLDs occurred more frequently in the initial than non-initial position of words and in longer rather than shorter words. Additionally, SLDs occurred more frequently in sentences that contained more ‘bunsetsu’ (a kind of linguistic unit in Japanese). Our study is the first to show that both word and sentence-level factors could contribute to SLDs in preschool children who stutter in agglutinating languages, such as Japanese. This aspect is rarely reported in psycholinguistic studies based on stuttering occurrence in inflecting languages, such as English.

Speech Interference as the Result of a Two-Pronged Negative Influence

The article considers the problem of the interaction of languages in the process of forming a bilingual personality and society. A comparative study of the material, the mother tongue and the studied language, reveals the causes of interference in bilingual speech in a foreign language. Traditionally, the causes of speech interference are determined by the characteristics of a native language, which are not found in the language studied, therefore they negatively affect the process of mastering the second language, which generates speech errors. This is a one-sided approach to understand interference in general and its origins (causes) in particular. The article considers the problem of a two-way approach to understanding the phenomenon of speech interference. Speech interference is a result of the negative influence of both the characteristics of the native language and the characteristics of the language being studied, i.e. it is a simultaneous two-way process in dual unity. Both processes of negative influence on an individuals mastery of a second language are defined as one action in bilinguality, giving the same result - interference in bilingual speech, which is expressed in phonetic-phonological, semantic and other types and types of speech errors. The simultaneous and inconsistent negative influence of the features of the native language and the non-native language considered on the material of consonant combinations in the initial words of the Russian and Kazakh languages, when the absence of combinations of consonants in this position of the Kazakh word and their presence in this position of the Russian word to the same extent and at the same time negatively influence on the correct - the literary pronunciation of Kazakh words and Russian words. Thus, in contrast to the traditional explanation of the phenomenon of speech interference as a result of a one-sided negative process, it is proposed and proved that this phenomenon is the result of (simultaneous) two-way influence of features, native and studied languages. Errors of an individual in speech in a foreign language are considered according to the provisions of the syntagmatic typology of interference (plus segmentation, or minus segmentation).

End-to-end event factuality prediction using directional labeled graph recurrent network

Event factuality prediction is the task of assessing the degree to which an event mentioned in a sentence has happened. However, existing methods usually stack encoders to make factuality predictions given the gold positions of anchor words. In addition, the frequently used encoders, such as bidirectional LSTMS and graph convolution networks, ignore the directional labeled syntactic information while modeling the context. To fill the gap when facing plain text without identifying event anchor words in advance, we investigate the task of end-to-end EFP in this paper. We present the Directional Labeled Graph Recurrent Network, denoted as DLGRN, to solve Event Anchor Detection and Factuality Induction in a multi-task framework. Specifically, we represent sentences as syntactic information graphs. Then, to incorporate directional labeled information, we design edge-tied weights and edge-aware attention mechanism on top of a graph-based recurrently message passing encoder. We further propose to utilize multi-task learning to jointly model Event Anchor Detection and Factuality Induction by optimizing a mixed-objective learning function. We use four widely used factuality prediction benchmarks (i.e., FactBank, Meantime, UW, and UDS-IH2) to evaluate our framework. Our framework achieves state-of-the-art results in the two subtasks, averagely decreasing 17.12% MAE and raising 5.40% Pearson correlation r against the best baseline. In addition, experimental results show that our framework can capture the overall factuality score distributions, and incorporating directional and labeled syntactic information in EFP achieves better performances than the baselines.

Consoantes róticas duplas em contexto inicial de palavra nas cantigas medievais: casos de hipossegmentação

This article reflects on the cases of hyposegmentation that involve double rhotic consonants in an initial context of words and clitics in Ancient Portuguese (AP). The objective of the study is to map, in the troubadour era, the hyposegmentation cases of the rhotic consonant in the initial word position. The mapping indicates a slight variation in the spelling of words that begin with this consonant and the lack of blanks is systematically found in the context of the presence of clitic followed by the studied consonant.

The production of ejectives in German and Georgian

This paper aims to shed light on the textbook description of ejective production through an analysis of the articulatory measures intraoral pressure and larynx height in a data set of phonological Georgian ejectives and epiphenomenal German ejectives. By comparing articulatory mechanisms for ejective production in these two languages, we aim to learn about how ejectives gain phonological status in a language. 14 female German speakers and 13 female Georgian speakers are recorded in a multi-channel experimental set-up. We collect intraoral pressure (IOP) data, an electroglottographic signal and larynx trace using a dual-channel electroglottograph, as well as audio data. We control for place of articulation, vocalic context, word and sentence position, as well as for plosive duration in our analysis. Generalized additive mixed-models are used for statistical analysis. In Georgian, larynx height highly depends on the word and sentence position of the ejective. We find evidence for the hypothesis that pressure buildup after oral closure may be realized by a pulmonic airstream in both Georgian and German ejective production. Sentence-internal Georgian ejectives are produced with lower IOP peak and less steep IOP increase than sentence-initial ejectives. Epiphenomenal ejectives in German are not produced with raised larynx and show lower average IOP than plosives in control conditions.

Mandarin adaptations of voiced alveolar fricative [z] in English loanwords

This paper explores the adaptation of English voiced fricative [z] E into Mandarin. The principal findings are that the adaptation of [z] E depends on its position in the source word. If [z] E occupies the word initial or middle position, it tends to be borrowed as Mandarin [ts] M or [s] M . If [z] E is in the ending position, more variations would be observed as it may correspond to [ts] M , [s] M , [ʈʂ] M or [ʂ] M . Moreover, adapting [z] E into Mandarin is also heavily influenced by the orthography. The English [z] E can be spelled with &lt;s&gt; or &lt;z&gt;. The letters of &lt;s&gt; and &lt;z&gt; are also listed in Mandarin Pinyin , a Romanization system that transcribes the sounds of Chinese characters using Roman alphabets. Specifically, &lt;s&gt; is pronounced as [s] M , while [z] is pronounced as [ts] M . It turns out that [s] M sounds different from [z] E , while [ts] M is perceptually similar to [z] E . Generally, there emerged two major adaptation patterns, that is, &lt;z&gt; [z] E " &lt;z&gt; [ts] M and &lt;s&gt; [z] E " &lt;s&gt; [s] M . The latter is based on the spelling similarity, rather than perceptual similarity. It is therefore concluded that the loanword adaptation is not only determined by the speech perception, phonology and legitimacy of sound structures, but also systematically interfered with the orthography, that is, the source-loan spelling similarity.

Word-Final Disfluencies in a School-Age Child: Beneath the Tip of the Iceberg.

Purpose Disfluencies associated with stuttering generally occur in the initial position of words. This study reviews data from a school-age child with an atypical stuttering profile consisting predominantly of word-final disfluencies (WFDs). Our primary goals were to identify patterns in overt features of WFDs and to extend our understanding of this clinical profile by focusing on aspects of stuttering that lie beneath the surface. Method Analyses explored the patterns and distributions of the child's observable stuttering behaviors, in addition to his awareness, perceptions, and subjective experience of stuttering. Results Findings indicated that the WFD profile consists of relatively consistent and distinct overt features. We also found that, in many ways, the child perceived the impact of his disfluencies much like other children who stutter, even though his disfluencies manifested in a less common form. Conclusion Findings suggest the need for increased awareness of the varied forms stuttering may take in order to ensure accurate diagnosis, clear communication to clients and parents, and timely access to appropriate intervention.

Acoustic characteristics of Korean-English bilingual speakers’ /l/ and the relationship to their foreign accent ratings.

PurposeThe current study examined acoustic characteristics of English /l/ produced by Korean speakers who speak English as their second language (KE). Comparisons were made to patterns of their Korean /l/ and English /l/ produced by monolingual speakers of American English (AE). Vowel context and word position effects on acoustic differences among the three /l/ sounds were examined in relation to the degree of foreign accentedness of KE speakers. MethodsKorean and English word-initial and -final /l/ in two vowel contexts (high front vs. low back) were produced by 12 KE female adult speakers in a carrier phrase. A comparison group of 10 AE speakers produced English /l/. F1 and F2 values of each /l/ production were extracted at the /l/ midpoint and were compared to each other. Euclidean distances across the productions were also examined in relation to the degrees of foreign accentedness of KE speakers. ResultsKE speakers’ English /l/ showed significantly lower F2-F1 values than their Korean /l/ across word position and vowel context but significantly higher F2-F1 values than AE speakers’ English word-initial /l/. Euclidean distance between KE speakers’ English and Korean /l/, and KE and AE speakers’ English /l/ correlated significantly with KE speakers’ degree of foreign accentedness across word position and vowel context. ConclusionResults indicate that the influence of Korean on KE speakers’ English /l/ productions varied by word position and vowel context, as well as the degree of their foreign accentedness ratings. This finding suggests the context-dependent nature of /l/ productions and highlights the need for considering a speakers’ degree of foreign accent when exploring English /l/ patterns of KE speakers.

MS-Transformer: Introduce multiple structural priors into a unified transformer for encoding sentences

Transformers have been widely utilized in recent NLP studies. Unlike CNNs or RNNs, the vanilla Transformer is position-insensitive, and thus is incapable of capturing the structural priors between sequences of words. Existing studies commonly apply one single mask strategy on Transformers for incorporating structural priors while failing at modeling more abundant structural information of texts. In this paper, we aim at introducing multiple types of structural priors into Transformers, proposing the Multiple Structural Priors Guided Transformer (MS-Transformer) that transforms different structural priors into different attention heads by using a novel multi-mask based multi-head attention mechanism. In particular, we integrate two categories of structural priors, including the sequential order and the relative position of words. For the purpose of capturing the latent hierarchical structure of the texts, we extract these information not only from the word contexts but also from the dependency syntax trees. Experimental results on three tasks show that MS-Transformer achieves significant improvements against other strong baselines.

Attention-based position-aware framework for aspect-based opinion mining using bidirectional long short-term memory

Aspect-based Opinion Mining is a form of fine-grained Sentiment Analysis and it models the semasiological relationship between aspect terms and context words in a sentence. The presence of a variety of context words has a significant impact on a sentence's sentiment polarity. As a result, while designing a model, it is necessary to consider the interaction of aspects and context words. Although existing approaches have taken into account an aspect’s position in a sentence, much of the research works have not explored the use of Sentiment Lexicons with the Deep Learning algorithms. In this paper, we propose a framework for an Attention-based position-aware Bidirectional Long Short-Term Memory network for Aspect-based Opinion Mining that incorporates a Sentiment Intensity Lexicon. The aspect word’s pre-trained vector is adjusted to be closer to semantically and sentimentally similar nearest neighbors and further away from sentimentally dissimilar neighbors. The proposed framework calculates aspect weights by concatenating the external knowledge in the form of lexicon sentiment intensity scores with word embeddings and position information. The framework experiments on the SemEval 2014 dataset. The results of the experiments illustrate that injecting external knowledge into the Bidirectional Long Short-Term Memory network can improve classification accuracy significantly.

ASMĀ' AL-A'DĀD WA KINĀYĀTUHA FĪ AL-QUR'AN AL-KARĪM (DIRĀSAH TAHLĪLIYYAH NAHWIYYAH)

In the types of words there is a chapter on numerals and their kinayah. It is widely used in Arabic, either in conversation or in Arabic texts. The purpose in this research is to describe the number words and their kinayah, as well as the form of the number word phrase and the position of the number word and its kinayah in the verse of the holy Quran. The type in this research is library research with a content analysis design (adapted from Krippendorf) because it is based on the data sources in this study in the form of documents. The primary data in this study was the holy Quran and the secondary data in this study were taken from books related to the research. And the results of this study, 1. Quantity of number word 37 words in the Quran 2. Quantity of kinayah number in the Quran 4 words 3. Quantity of position number 17 verse in the Quran 4. Quantity of position kinayah number 6 verse in the Quran.

The supralaryngeal articulation of stress and accent in Greek

It is well reported that articulatory movements comprising prominence units are longer, larger, and faster than their non-prominent counterparts. However, it is unclear whether these effects arise at the level of lexical stress, accent, or both, reflecting a hierarchy of prominence, i.e., being stronger when induced by accent as opposed to stress. It is also uncertain whether prominence-induced kinematic effects are invariant across positions of stress within the word, types of focus that accent denotes, and positions of words in the phrase. Here, we use an electromagnetic articulography (EMA) study to assess the supralaryngeal kinematic correlates of prominence in Greek across three stress positions (antepenultimate, penultimate, ultimate; i.e., all possible stress positions in Greek), two accentual conditions (accented and de-accented), and two phrasal positions (phrase-medial and phrase-final). Focus type is also considered, with the accentual conditions coming from two types of focus (broad and narrow), while the de-accented conditions are by default unfocused. Our results indicate that stressed syllables involve longer, larger, and faster gestures than their unstressed counterparts, regardless of the position of stress within the word. Notably, variation in velocity is accounted for by variation in displacement. Presence of accent does not further expand the stressed gestures, although it is related to minimal kinematic changes across the whole word, the exact profile of which depends on stress position. With the exception of final vowel duration, focus type is not systematically encoded in these kinematic effects. Finally, interactions are detected between the kinematic profile of prominence and that of boundaries. Implications of our findings for the hierarchy of prominence and its cross-linguistic differences are discussed, and a gestural account of prominence and boundaries is put forward.

Glimpsing keywords across sentences in noise: A microstructural analysis of acoustic, lexical, and listener factors.

This study investigated how acoustic and lexical word-level factors and listener-level factors of auditory thresholds and cognitive-linguistic processing contribute to the microstructure of sentence recognition in unmodulated and speech-modulated noise. The modulation depth of the modulated masker was changed by expanding and compressing the temporal envelope to control glimpsing opportunities. Younger adults with normal hearing (YNH) and older adults with normal and impaired hearing were tested. A second group of YNH was tested under acoustically identical conditions to the hearing-impaired group, who received spectral shaping. For all of the groups, speech recognition declined and masking release increased for later keywords in the sentence, which is consistent with the word position decreases in the signal-to-noise ratio. The acoustic glimpse proportion and lexical word frequency of individual keywords predicted recognition under different noise conditions. For the older adults, better auditory thresholds and better working memory abilities facilitated sentence recognition. Vocabulary knowledge contributed more to sentence recognition for younger than for older adults. These results demonstrate that acoustic and lexical factors contribute to the recognition of individual words within a sentence, but relative contributions vary based on the noise modulation characteristics. Taken together, acoustic, lexical, and listener factors contribute to how individuals recognize keywords during sentences.

Acoustic Characteristics of Pre- and Post-vocalic /l/: Patterns from One Southern White Vernacular English.

This study examined acoustic characteristics of the phoneme /l/ produced by young female and male adult speakers of Southern White Vernacular English (SWVE) from Louisiana. F1, F2, and F2-F1 values extracted at the /l/ midpoint were analyzed by word position (pre- vs. post-vocalic) and vowel contexts (/i, ɪ/ vs. /ɔ, a/). Descriptive analysis showed that SWVE /l/ exhibited characteristics of the dark /l/ variant. The formant patterns of /l/, however, differed significantly by word position and vowel context, with pre-vocalic /l/ showing significantly higher F2-F1 values than post-vocalic /l/, and /l/ in the high front vowel context showing significantly higher F2-F1 values than those in the low back vowel context. Individual variation in the effects of word position and vowel contexts on /l/ pattern was also observed. Overall, the findings of the current study showed a gradient nature of SWVE /l/ variants whose F2-F1 patterns generally fell into the range of the dark /l/ variant, while varying by word position and vowel context.