学科服务 更多>>

关键词： precipitation verification   machine learning   categorical indexes   neural networks   modeling  

摘要： This work introduces a methodology for optimizing neural network models using a combination of continuous and categorical binary indices in the context of precipitation forecasting. Probability of detection and false alarm rate are popular metrics used in the verification of precipitation models. However, machine learning models trained using gradient descent cannot be optimized based on these metrics, as they are not differentiable. We propose an alternative formulation for these categorical indices that are differentiable and we demonstrate how they can be used to optimize the skill of precipitation neural network models defined as a multiobjective optimization problem. To our knowledge, this is the first proposal of a methodology for optimizing weather neural network models based on categorical indices.

关键词： category interest   curiosity driven learning   pupil dilation   word learning   word recognition   vocabulary size  

摘要： The overall pattern of vocabulary development is relatively similar across children learning different languages. However, there are considerable differences in the words known to individual children. Historically, this variability has been explained in terms of differences in the input. Here, we examine the alternate possibility that children's individual interest in specific natural categories shapes the words they are likely to learn - a child who is more interested in animals will learn a new animal name easier relative to a new vehicle name. Two-year-old German-learning children (N = 39) were exposed to four novel word-object associations for objects from four different categories. Prior to the word learning task, we measured their interest in the categories that the objects belonged to. Our measure was pupillary change following exposure to familiar objects from these four categories, with increased pupillary change interpreted as increased interest in that category. Children showed more robust learning of word-object associations from categories they were more interested in relative to categories they were less interested in. We further found that interest in the novel objects themselves influenced learning, with distinct influences of both category interest and object interest on learning. These results suggest that children's interest in different natural categories shapes their word learning. This provides evidence for the strikingly intuitive possibility that a child who is more interested in animals will learn novel animal names easier than a child who is more interested in vehicles.

关键词： Phoneme perception   Phonological representations   Phonetic ambiguity   Lexical access  

摘要： This study examines ERP correlates of the different processes associating two phones to one vs. two phonemic categories in two regional varieties of French. Two groups of French listeners are compared, respectively exploiting two regional varieties, with a contrast between the mid-low /e/ and the mid-high /e/ for Northern French (NF) but not for Southern French (SF). It is expected that the competition between the two close categories /e/ vs. /e/ in NF could induce an ERP modulation in the processing of/e/compared to a phoneme/a/with no close competitor, serving as control. In contrast, there should be no difference in ERP response in /e/ vs. /a/ in SF, where there is no competition between the close neighbors /e/ and /e/. The participants of the two groups listened to words containing either /e/ or /a/ in a go/no-go semantic categorization task in which the critical /e/ and /a/ words did not require an overt behavioral response. We found a significant difference in the N400 amplitude between the two conditions in the NF but not in the SF variety. The fact that the ERP modulations appear on the N400 component suggests that lexical access is influenced by the regional variety of the speakers.

关键词： cardiovascular risk factors   obesity  

摘要： Background Cardiovascular disease (CVD) is a significant cause of morbidity and mortality globally. Obesity is an important CVD risk factor and is increasing in prevalence. Methods In this study, 3829 men and 5720 women (35-65 years) were enrolled as part of the MASHAD cohort study. Four categories were identified according to body mass index and waist circumference that was defined by the World Health Organization. Logistic regression analysis was used to determine the adjusted odds ratio (OR) for the occurrence of CVD, and Cox regression model was used to evaluate the association of obesity with CVD incidence. Results We found that the higher risk groups defined by categories of adiposity were significantly related to a higher prevalence of a high serum total cholesterol (TC), and triglycerides (TG), and lower high-density lipoprotein cholesterol (HDL), and higher fasting blood glucose (FBG) in both genders and a higher low-density lipoprotein cholesterol (LDL) in women (P < .001). Additionally, a high percentage of participants with dyslipidemia, high LDL, high TC, and low HDL and a high percentage of participants with metabolic syndrome, diabetes, hypertension, and a high serum TG were observed across obesity categories (P ). Moreover, women with the very high degrees of obesity had a greater risk of CVD (HR: 1.91, 95% CI: 1.06-3.43, P = .03). Conclusion Obesity strongly predicts several CVD risk factors. Following 6 years of follow-up, in individuals within increasing degrees of obesity, there was a corresponding significant increase in CVD events, rising to approximately a twofold higher risk of cardiovascular events in women compared with men.

关键词： almost omega-categoricity   weak o-minimality   quite o-minimality   binary theory  

摘要： Studying the properties of almost omega-categorical quite o-minimal theories, we prove that the arbitrary families of pairwise weakly orthogonal nonalgebraic 1-types in these theories are orthogonal. We also establish the binarity of these theories.

关键词： Group Lasso   Regularization   Standardization   Variable Selection  

摘要： Categorical regressor variables are usually handled by introducing a set of indicator variables, and imposing a linear constraint to ensure identifiability in the presence of an intercept, or equivalently, using one of various coding schemes. As proposed in Yuan and Lin (2006), the group lasso is a natural and computationally convenient approach to perform variable selection in settings with categorical covariates. As pointed out by Simon and Tibshirani (2012), "standardization" by means of block-wise orthonormalization of column submatrices each corresponding to one group of variables can substantially boost performance. In this note, we study the aspect of standardization for the special case of categorical predictors in detail. The main result is that orthonormalization is not required;column-wise scaling of the design matrix followed by re-scaling and centering of the coefficients is shown to have exactly the same effect. Similar reductions can be achieved in the case of interactions. The extension to the so-called sparse group lasso, which additionally promotes within-group sparsity, is considered as well. The importance of proper standardization is illustrated via simulations and a case study. (C) 2019 Elsevier B.V. All rights reserved.

关键词： ATC   Anatomical Therapeutic Chemical   AWaRe   Access, Watch, and Reserve   URTI   Upper respiratory tract infection   WHO   World Health Organization  

摘要： Objectives To assess clinical indication-specific antibiotic prescribing in pediatric practice in China based on the World Health Organization (WHO) Access, Watch, and Reserve (AWaRe) metrics and to detect potential problem areas. Study design Pediatric prescription records on the 16th of each month during 2018 were sampled for all encounters at outpatient and emergency departments of 16 tertiary care hospitals via hospital information systems. Antibiotic prescribing patterns were analyzed across and within diagnostic conditions according to WHO AWaRe metrics and Anatomical Therapeutic Chemical (ATC) classification. Results A total of 260 001 pediatric encounters were assessed, and antibiotics were prescribed in 94 453 (36.3%). In 35 167 encounters (37.2%), at least 1 intravenous antibiotic was administered. WHO Watch group antibiotics accounted for 82.2% (n = 84 176) of all antibiotic therapies. Azithromycin (n = 15 791;15.4%) was the most commonly prescribed antibiotic, and third-generation cephalosporins (n = 44 387;43.3%) were the most commonly prescribed antibiotic class. In at least 66 098 encounters (70.0%), antibiotics were prescribed for respiratory tract conditions, mainly for bronchitis/bronchiolitis (n = 25 815;27.3%), upper respiratory tract infection (n = 25 184;26.7%), and pneumonia (n = 13 392;14.2%). Conclusions Overuse and misuse of WHO Watch group antibiotics for respiratory tract conditions and viral infectious diseases is common in pediatric outpatients in China. Pediatric antimicrobial stewardship should be strengthened using WHO AWaRe metrics.

关键词： Natural Monopoly   Brand loyalty   Heavy buyers   Light buyers   Store brands   Manufacturer brands  

摘要： This paper investigates the Natural Monopoly [NM] effect, which is that large brands have buyers who are on average less frequent or 'lighter' purchasers of the product category. The study analyzes the NM effect for brands in 28 consumer goods categories in The Netherlands. The analysis employs a multiple regression with category purchase rate as the dependent variable;and brand penetration, together with brand price, brand type, average pack size and promotion incidence as independent variables. The study finds that higher brand penetration is indeed associated with a lower rate of category purchase, controlling for the other variables in the model. The NM effect is reasonably large: the largest two brands in a category tend to have a buyer base that on average purchase the category about 25% less frequently than those of the smallest two. The study also derives an explanation for how large brands are generally purchased more frequently, even when their buyer base on average buys the category less frequently. The findings imply that a focus on heavy category buyers is inconsistent with the goal of growing a brand. (C) 2020 Australian and New Zealand Marketing Academy. Published by Elsevier Ltd. All rights reserved.

关键词： arthritis   biomarkers   childhood arthritis   cluster analysis   cytokines   data mining   juvenile idiopathic arthritis  

摘要： Objective. To identify discrete clusters comprising clinical features and inflammatory biomarkers in children with JIA and to determine cluster alignment with JIA categories. Methods. A Canadian prospective inception cohort comprising 150 children with JIA was evaluated at baseline (visit 1) and after six months (visit 2). Data included clinical manifestations and inflammation-related biomarkers. Probabilistic principal component analysis identified sets of composite variables, or principal components, from 191 original variables. To discern new clinical-biomarker clusters (clusters), Gaussian mixture models were fit to the data. Newly-defined clusters and JIA categories were compared. Agreement between the two was assessed using Kruskal-Wallis analyses and contingency plots. Results. Three principal components recovered 35% (three clusters) and 40% (five clusters) of the variance in patient profiles in visits 1 and 2, respectively. None of the clusters aligned precisely with any of the seven JIA categories but rather spanned multiple categories. Results demonstrated that the newly defined clinical-biomarker lustres are more homogeneous than JIA categories. Conclusion. Applying unsupervised data mining to clinical and inflammatory biomarker data discerns discrete clusters that intersect multiple JIA categories. Results suggest that certain groups of patients within different JIA categories are more aligned pathobiologically than their separate clinical categorizations suggest. Applying data mining analyses to complex datasets can generate insights into JIA pathogenesis and could contribute to biologically based refinements in JIA classification.

关键词： Health-related quality of life   Minnesota Living with Heart Failure Questionnaire   Item characteristic curves   Survey response options   Heart failure  

摘要： Purpose Half of the 21-item Minnesota Living with Heart Failure Questionnaire (MLHFQ) response categories are labeled (0 = No, 1 = Very little, 5 = Very much) and half are not (2, 3, and 4). We hypothesized that the unlabeled response options would not be more likely to be chosen at some place along the scale continuum than other response options and, therefore, not satisfy the monotonicity assumption of simple-summated scoring. Methods We performed exploratory and confirmatory factor analyses of the MLHFQ items in a sample of 1437 adults in the Better Effectiveness After Transition-Heart Failure study. We evaluated the unlabeled response options using item characteristic curves from item response theory-graded response models for MLHFQ physical and emotional health scales. Then, we examined the impact of collapsing response options on correlations of scale scores with other variables. Results The sample was 46% female;71% aged 65 or older;11% Hispanic, 22% Black, 54% White, and 12% other. The unlabeled response options were rarely chosen. The standard approach to scoring and scores obtained by collapsing adjacent response categories yielded similar associations with other variables, indicating that the existing response options are problematic. Conclusions The unlabeled MLHFQ response options do not meet the assumptions of simple-summated scoring. Further assessment of the performance of the unlabeled response options and evaluation of alternative scoring approaches is recommended. Adding labels for response options in future administrations of the MLHFQ should be considered.