学科服务 更多>>

关键词： Feature screening   Partition-based   Categorical data   RKHS  

摘要： This paper proposes a new screening procedure for the ultrahigh dimensional data with a categorical response. By exploiting the group structure among predictors, a new partition-based screening approach is developed via the reproducing kernel Hilbert space (RKHS) embeddings in the maximum mean discrepancy framework. Consequently, the new method is able to identify the influential group of predictors that may be overlooked by the marginal screening methods. Moreover, by using the RKHS embedding, the new ranking index has a very simple form, and thus can be evaluated easily. As a byproduct, the new method is model-free without specifying any relationship between the predictors and the response. The sure screening property of the proposed method is proved and the effectiveness of the new method is also illustrated via numerical studies and a real data analysis. (C) 2021 Elsevier B.V. All rights reserved.

关键词： atypical design   category prestige   package design  

摘要： Innovation in the design of products has become an important competitive tool of success in the market for many firms (see the virtual issue on Design Innovation in the Journal of Product Innovation Management [2020]). A few firms have had success in the market by adopting innovative designs. But not all innovative designs give success to firms. This research focuses on package design as an important marketing tool of innovation, influencing consumer responses to products. When designing packaging, firms typically follow category-based visual codes, but they can adopt an innovative design that deviates from the codes, producing atypical package designs. This research shows that the effectiveness of atypical package design depends on product category prestige. Consumers respond to signals about product quality, such as package design, when they believe the signals are credible. When a product category is prestigious, consumers perceive package design as a diagnostic signal. Consumers are more likely to infer high product performance from an atypical package design and choose that product. We compare three dimensions of product design (aesthetics, functionality, and symbolism) to reveal that functionality is the reason why consumers prefer an atypical package design in a prestigious category. A series of studies show that the preference for atypical package design becomes stronger when consumers are choosing products in a prestigious category than in a standard category. The preference for atypical package design is attenuated when other information related to performance is provided, even when the aesthetics and symbolism of atypical package design do not change.

关键词： attention   category contrast   invention   patents   search  

摘要： Research Summary Whereas prior innovation and strategy literature studied how attentional and search dynamics influence the creation of inventions, we examine how these same processes affect the impact of inventions after their creation. We theorize that inventions classified in "high-contrast" technological categories garner more attention by potential users and, hence, accrue more citations than otherwise-equivalent inventions classified in "low-contrast" categories. We test this hypothesis via three studies. First, we estimate citation-count models among all USPTO patents granted between 1975 and 2010. Second, we conduct a twin patents test comparing inventions patented both at the USPTO and at the EPO. Third, we examine minute-by-minute search logs from a sample of USPTO examiners. These studies support our hypothesis and extend current understandings of attentional and search dynamics in the innovation process. Managerial Summary Patents that receive more citations tend to have greater economic value and greater impact on future technological developments. We show that the number of citations a patent receives does not only depend on its inherent technological value, but also on seemingly neutral classification decisions affecting the likelihood that it will be noticed by potential future users. We test our arguments via three related studies. Our results demonstrate that inventions classified in "high-contrast" technology classes garner considerably more attention-and hence citations-than twin-inventions classified in "low-contrast" classes. The key managerial implication is that, whenever feasible, nudging an invention towards higher-contrast classes will increase its future worth. The key policy implication is that maximizing categorical contrast across technology classes will help users identify relevant prior patents.

关键词： DNA function   deep neural network   category attention  

摘要： Quantifying DNA properties is a challenging task in the broad field of human genomics. Since the vast majority of non-coding DNA is still poorly understood in terms of function, this task is particularly important to have enormous benefit for biology research. Various DNA sequences should have a great variety of representations, and specific functions may focus on corresponding features in the front part of learning model. Currently, however, for multi-class prediction of non-coding DNA regulatory functions, most powerful predictive models do not have appropriate feature extraction and selection approaches for specific functional effects, so that it is difficult to gain a better insight into their internal correlations. Hence, we design a category attention layer and category dense layer in order to select efficient features and distinguish different DNA functions. In this study, we propose a hybrid deep neural network method, called DeepATT, for identifying 919 regulatory functions on nearly 5 million DNA sequences. Our model has four built-in neural network constructions: convolution layer captures regulatory motifs, recurrent layer captures a regulatory grammar, category attention layer selects corresponding valid features for different functions and category dense layer classifies predictive labels with selected features of regulatory functions. Importantly, we compare our novel method, DeepATT, with existing outstanding prediction tools, DeepSEA and DanQ. DeepATT performs significantly better than other existing tools for identifying DNA functions, at least increasing 1.6% area under precision recall. Furthermore, we can mine the important correlation among different DNA functions according to the category attention module. Moreover, our novel model can greatly reduce the number of parameters by the mechanism of attention and locally connected, on the basis of ensuring accuracy.

关键词： Variability   Category learning   Development   Neural adaptation   FMRI  

摘要： Perceptual variability is often viewed as having multiple benefits in object learning and categorization. Despite the abundant results demonstrating benefits such as increased transfer of knowledge, the neural mechanisms underlying variability as well as the developmental trajectories of how variability precipitates changes to category boundaries are unknown. By manipulating an individual's exposure to variability of novel, metrically organized categories during an fMRI-adaptation paradigm, we were able to assess the functional differences between similarity and variability in category learning and generalization across two time-points in development: adulthood (n = 14) and late childhood (n = 13). During this study, participants were repeatedly exposed to category members from different distributions. After a period of adaptation, a deviant stimulus that differed from the expected distribution was then presented. This deviant differed in either an invariant dimension (a feature that remained consistent throughout presentation was altered) or a similarity dimension (a feature that changed throughout exposure was changed in a new dimension). Our results can be summarized in three main findings: (1) Variability during exposure recruited the right fusiform gyrus to a greater extent than tight exposure. (2) Deviant items were generalized based on the exemplar distributions during exposure, although children only generalized items if provided variable exposure. (3) Variability influenced release to a greater extent in children than adults. These results are discussed in relation to the variability and category learning literature more broadly.

关键词： Branding   Consumers   Retail brand image   New retail products   Image inheritance  

摘要： Retailer brand collaborations are an underexploited way for retailers to expand product lines and target new segments relatively quickly and cheaply, yet little work has explored the area and the influence of important factors in the image inheritance process remains unknown. Using data from two experiments involving 240 subjects, we examine the role of brand type, brand strength and the fit between parent and brand collaboration product categories to show that a new retail product?s image inherits more characteristics from a symbolic parent brand (even when the brand is weak) and when there are higher degrees of parent retailer-image fit. However, fit between the new retail collaboration?s product category and that of the parent brands is only important for functional brands. The results help retailers decide which partner to choose to maximize image inheritance.

关键词： FI-modules   Torsion theories  

摘要： We work with FI-modules over a small preadditive category R, viewed as a ring with several objects. Our aim is to study torsion theories for FI-modules. We are especially interested in torsion theories on finitely generated FI-modules and the category of what we call "shift finitely generated" FI-modules. We also apply these methods to study inductive descriptions of FI-modules over R. (C) 2020 Elsevier B.V. All rights reserved.

关键词： Social Life Cycle Assessment   Local community   Participatory approach   Impact subcategory selection   Goal and scope   Multiple criteria decision analysis  

摘要： Purpose Standardized and generalized impact subcategories may not be representative of the social issues associated with the product analyzed in social life cycle assessment (S-LCA) studies. The analyst's knowledge about the product studied is currently considered to define the impact subcategories included in the study, characterizing a limitation to be overcome. This research proposes a method able to model the subjectivity associated with the selection of impact subcategories for S-LCA through a stakeholders' participatory approach. Methods This framework is composed of three stages and was applied for the stakeholder "Local Community" of the fruit pulp produced in northeastern Brazil. The first phase identifies the representatives of local communities around the product's life cycle and the potential subcategories that can be considered. In the second phase, the preference elicitation process of these representatives is carried out. In the third phase, the information obtained is aggregated to classify the impact subcategories into relevant and non-relevant through the PROMSORT method. Results and discussion The methodological steps were applied into two contexts: (i) individual perception and (ii) consensus perception. In the first case, the representatives considered the following relevant subcategories: access to material resources, cultural heritage, access to immaterial resources, and safe and healthy living conditions. However, the subcategories of community engagement and displacement and migration were not classified as strongly relevant. Only the subcategory respecting the rights of native peoples was considered non-pertinent. For the second case, the pertinent subcategories were access to material resources, cultural heritage, access to immaterial resources, displacement and migration, safe and healthy living conditions, and community engagement. The non-relevant subcategory also refers to respect for the rights of native peoples. Conclusions The use o

关键词： Higher Auslander algebra   d-Abelian category   d-Cluster tilting subcategory   Higher homological algebra   Wide subcategory  

摘要： A subcategory W of an abelian category is called wide if it is closed under kernels, cokernels, and extensions. Wide subcategories are of interest in representation theory because of their links to other homological and combinatorial objects, established among others by Ingalls-Thomas and Marks-aoviCek. If d >= 1 is an integer, then Jasso introduced the notion of d-abelian categories, where kernels, cokernels, and extensions have been replaced by longer complexes. Wide subcategories can be generalised to this situation. Important examples of d-abelian categories arise as the d-cluster tilting subcategories M-n,(d) of mod A(n)(d-1), where A(n)(d-1 )is a higher Auslander algebra of type A in the sense of Iyama. This paper gives a combinatorial description of the wide subcategories of M-n,M-d in terms of what we call non-interlacing collections. (C) 2020 Elsevier B.V. All rights reserved.