教学课程资源库 更多>>

关键词： Banzhaf value   Choquet integral   extended Archimedean Copula and extended Archimedean Co‐   Copula   group decision‐   making   information fusion   multigranular generalized orthopair 2‐   tuple linguistic information model  

摘要： In multiattribute group decision-making (MAGDM), experts often articulate their preference information to support decision-making by applying the multigranular linguistic model. Thus, the present work aims to introduce a novel MAGDM model to manage multigranular generalized orthopair 2-tuple linguistic information (GO2TLI). To begin with, a generalized orthopair 2-tuple linguistic model is put forward with the attempt of taking advantages of both q-rung orthopair fuzzy set (q-ROFS) and the 2-tuple linguistic model, while a transformation approach is supplied to tackle the consistency of multigranular GO2TLI. In addition, the Archimedean Copula as well as the Co-Copula operators are extended to handle GO2TLI along with their operational laws, to comprehensively model the relationship among attributes and experts. The Banzhaf Choquet-Copula aggregation operators on the generalized orthopair 2-tuple linguistic (GO2TLBCCA) are introduced, also some of its properties discussed. Third, the algorithms for regulating and determining the fuzzy measure (FM) of attributes and experts sets are proposed, followed by the corresponding decision-making approaches based upon the proposed GO2TLBCCA. The proposed MAGDM model can not only accommodate effectively the FMs of attribute (and expert) sets which are given subjectively, but also effectively address some multigranular GO2TLI as well as the partially unknown or completely unknown weights of attribute and expert sets. Finally, a case study is provided to demonstrate the validity of the proposed approach along with relevant discussions, the merits of the proposed approach are also analyzed by comparing with some extant decision methods.

关键词： Phonotactics   Demographics   Consonant clusters   Network structure  

摘要： There is an ongoing debate as to whether linguistic structure is influenced by demographic factors. Relationships between these two domains have been investigated on the phonological, morphological and lexical level, mainly drawing on synchronic data and comparative methodology. In this exploratory study, by contrast, we focus on the lesser recognized level of phonotactics, and adopt a methodologically orthogonal approach. We investigate the diachronic development of a single lineage, namely English, and compare it with concomitant developments of the demography of the English-speaking population. In addition to linguistic and demographic factors, we also derive characteristics of the underlying speaker network (network diameter;clustering coefficient). Empirically, we focus on the system of English consonant clusters, which we argue to be particularly sensitive to linguistic change so that effects of demography are expected to be more clearly visible than in more robust linguistic subsystems (e.g. phoneme inventory;morphology). By employing time-series clustering, it is shown that the trajectory of phonotactic diversity in English coda clusters most closely matches that of covariates related with density and heterogeneity of the speaker population. Linguistic covariates are less closely related. We conclude that heterogeneity of the linguistic input and the number of informants a learner is exposed to are relevant factors in the evolution of phonotactic diversity. (c) 2021 Elsevier Ltd. All rights reserved.

关键词： machine translation   neural network   morphology   vocabulary  

摘要： Neural Networks applied to Machine Translation need a finite vocabulary to express textual information as a sequence of discrete tokens. The currently dominant subword vocabularies exploit statistically-discovered common parts of words to achieve the flexibility of character-based vocabularies without delegating the whole learning of word formation to the neural network. However, they trade this for the inability to apply word-level token associations, which limits their use in semantically-rich areas and prevents some transfer learning approaches e.g. cross-lingual pretrained embeddings, and reduces their interpretability. In this work, we propose new hybrid linguistically-grounded vocabulary definition strategies that keep both the advantages of subword vocabularies and the word-level associations, enabling neural networks to profit from the derived benefits. We test the proposed approaches in both morphologically rich and poor languages, showing that, for the former, the quality in the translation of out-of-domain texts is improved with respect to a strong subword baseline.

摘要： From a socio-cultural perspective, language offers a means for children to communicate with and learn from others through interaction: language is the medium through which young children are provided cognitive, social, and emotional support in interactions with caregivers, siblings, and peers;and children characterized as dual language learners (DLLs) have in common that they receive this developmental support in two different languages. However, due to variations in socioeconomic factors, ethnic/immigration background, and language socialization practices, DLLs display considerable variability in their first- and second-language proficiency (McCabe, Tamis-LeMonda, Bornstein, Cates, Golinkoff, Guerra, Hirsh-Pasek, Hoff, Kuchirko, Melzi, Mendelsohn, Paez & Song, 2013).

摘要： This commentary responds to Meng Huat Chau and George M. Jacobs's article "Applied Linguistics, language guidelines, and inclusive practices: The case for the use of who with nonhuman animals" (same journal issue).

关键词： Linguistic Proximity Model   L3 acquisition   Non-facilitation   Transfer  

摘要： When we think of the debates surrounding linguistic transfer in L3 acquisition, one of the most prominent discussions concerns whether transfer occurs in a wholesale fashion or whether it is property-by-property. One such model is the Linguistic Proximity Model (LPM, Mykhaylyk et al., 2015;Westergaard et al., 2017;Westergaard, 2019), which maintains that transfer is property-by-property, with what Westergaard refers to as Full Transfer Potential (FTP). Westergaard injects the notion of complexity at each stage of development and recognizes the need to determine how a range of variables drive outcomes across these different stages. With that said, there are a set of points in the proposal that we believe are short of explanatory logic and will benefit from further consideration;we focus on two here. The first regards the need to go beyond post-hoc explanations of non-facilitative transfer via a commitment to a testable, proposal for when the LPM predicts such transfer will occur. The second relates to the current trend of using existing data to support property-by-property versus wholesale transfer. We contend that this application of existing data is an unsound practice because these data are in fact compatible with multiple theoretical accounts.

关键词： CALD   Culturally adapted   Health education   Chronic illness  

摘要： Objective: To review the effectiveness of health education interventions adapted for culturally and linguistically diverse (CALD) populations with a chronic illness. Methods: A systematic review and meta-analysis were conducted. Eligible studies were identified across six databases. Data were extracted and intervention effect was summarized using standardized mean difference. If there were insufficient data for meta-analysis, a descriptive summary was included. Modifying effects of intervention format, length, intensity, provider, self-management skills taught, and behavioral change techniques (BCTs) utilized were examined. Results: 58 studies were reviewed and data were extracted for 36 outcomes. Most interventions used multiple modes of delivery and were facilitated by bilingual health care professionals (HCPs). On average, interventions included 5.19 self-management skills and 4.82 BCTs. Interventions were effective in reducing BMI, cholesterol, triglycerides, blood glucose, HbA1C, and depression, and in increasing knowledge. Effectiveness was influenced partly by provider, with HCPs favored over lay providers or paraprofessionals in increasing knowledge;however, the opposite was noted for HbA1c. Conclusions: Health education interventions are effective among CALD populations, particularly at improving objective, distal outcomes (e.g., anthropometric measures). These interventions may be equally effective in improving proximal patient-reported outcomes (PROs);however, diversity in PROs limited analyses. Practice implications: Core outcome sets (COS) are needed to further investigate and compare health education intervention effectiveness on PROs. (c) 2021 Elsevier B.V. All rights reserved.

关键词： Kenya   Autosomal STRs   Population structure   Linguistic structure   GlobalFiler  

摘要： Kenya is a diverse and populous nation that employs DNA evidence in its criminal justice system, and therefore requires reliable information on autosomal STR allele frequency variation across the country and in its many ethnic groups. In order to provide reference data and to assess population structure, we analysed the 21 autosomal STRs in the GlobalFiler multiplex in a sample of 510 indigenous Kenyans representing the country's eight former provinces, 43 of its 47 counties, three main linguistic families and all 29 ethnic groups that each comprise >0.5% of the 2019 census population. The indigenous population originated from successive migrations of Cushitic, Nilotic and Bantu speaking groups who settled in regions that suited their distinctive sustenance lifestyles. Consequently, they now largely reside in a patchwork of communities with strong associations with particular counties and provinces and limited degrees of inter-group marriage, as shown by DNA donors' ancestry details. We found significant genetic differentiation between the three Nilotic language sub-families, with Western Nilotes (the Luo ethnic group) showing greater similarity to the Bantu than the Southern and Eastern Nilotes which themselves showed closer affinity to the Cushitic speakers. This concurs with previous genetic, linguistic and social studies. Comparisons with other African populations also showed that linguistic affiliation is a stronger factor than geography. This study revealed several rare off-ladder alleles whose structure was determined by Sanger sequencing. Among the unusual features that could affect profile interpretation were a deletion of Amelogenin Y but no other forensic marker (autosomal or Y-chromosomal), a triallelic pattern at TPOX and an extremely short SE33 allele falling within the expected size range of D7S820. Compared with the currently implemented Identifiler multiplex, Random Match Probabilities decreased from 6.4 x 10-19 to 3.9 x 10-27. The appreciation of l