摘要:Depression, as a common chronic mental disorder, has complex causes and low recovery rates. A method for assisting the diagnosis of depression using a hybrid bilinear deep learning network based on scalp electroencephalography is proposed. Firstly, the spatial features extracted by the convolutional neural network and the spatiotemporal features extracted by the convolutional long short-term memory network are fused into second-order hybrid features through bilinear methods to construct a hybrid bilinear model. Then, the functional connectivity matrices of each frequency band of EEG signals are used to train the model, and different functional connectivity measurement methods are used to analyze the relationship between the functional connectivity of each frequency band of EEG signals and depression. Finally, this method is applied on the MODMA dataset. The experimental results showed that the hybrid bilinear model using second-order hybrid features achieved an accuracy of 99.38% on the Beta frequency band correlation functional connectivity matrix, which indicates the effectiveness of the second-order hybrid features of the Beta frequency band correlation functional connectivity matrix in the auxiliary diagnosis of depression. Compared with other methods, the proposed method achieves higher accuracy and has high application prospects.
关键词:depression;functional connection;convolutional long-short term memory;bilinear;second-order
摘要:Multimodal sentiment analysis refers to recognizing the emotions expressed by characters in a video through textual, visual and acoustic information. Most of the existing methods learn multimodal coherence information by designing complex fusion schemes, while ignoring inter-and intra-modal differentiation information, resulting in a lack of information complementary to multimodal fusion representations. To this end, we propose a multi-subspace Transformer fusion network for multimodal sentiment analysis (MSTFN) method. The method maps different modalities to private and shared subspaces to obtain private and shared representations of different modalities, learning differentiated and unified information for each modality. Specifically, the initial feature representations of each modality are first mapped to their respective private and shared subspaces to learn the private representation containing unique information and the shared representation containing unified information in each modality. Second, under the premise of strengthening the roles of textual and audio modalities, a binary collaborative attention cross-modal Transformer module is designed to obtain textual and audio-based tri-modal representations. Then, the final representation of each modality is generated using modal private and shared representations and fused two by two to obtain a bimodal representation to further complement the information of the multimodal fusion representation. Finally, the unimodal representation, bimodal representation, and trimodal representation are stitched together as the final multimodal feature for sentiment prediction. Experimental results on two benchmark multimodal sentiment analysis datasets show that the present method improves on the binary classification accuracy metrics by 0.025 6/0.014 3 and 0.000 7/0.002 3, respectively, compared to the best benchmark method.
摘要:Facial expression recognition is an important research domain in psychology that can be applied to many fields such as transportation, medical care, security, and criminal investigation. Given the limitations of convolutional neural networks (CNN) in extracting global features of facial expressions, this paper proposes a Swin Transformer method embedded with a hybrid attention mechanism for facial expression recognition. Using the Swin Transformer as the backbone network, a hybrid attention module is embedded in the fusion layer (Patch Merging) in the model of Stage3, which can effectively extract global and local features from facial expressions.Firstly, the hierarchical Swin Transformer model can effectively obtain deep global features.Secondly, the embedded hybrid attention module combines channel and spatial attention mechanisms to extract features in the channel dimension and spatial dimension, which can attain better local features. At the same time, this article uses the transfer learning method to initialize the model network weights, thereby improving the recognition performance and generalization ability.The proposed method achieved recognition accuracies of 73.63%, 87.01%, and 98.28% on three public datasets (FER2013, RAF-DB, and JAFFE)respectively, achieving good recognition results.
摘要:Aiming at the problems of low recognition accuracy and poor generalization ability of current multimodal emotion recognition algorithms in modal feature extraction and information fusion between modalities, a multimodal emotion recognition algorithm based on speech, text and expression is proposed. Firstly, a shallow feature extraction network (Sfen) combined with parallel convolution module (Pconv) is designed to extract the emotional features in speech and text. A modified Inception-ResnetV2 model is adopted to capture the emotional features of expression in video stream. Secondly, in order to strengthen the correlation among modalities, a cross attention module is designed to optimize the fusion between speech and text modalities. Finally, a bidirectional long and short-term memory module based on attention mechanism (BiLSTM-Attention) is used to focus on key information and maintain the temporal correlation between modalities. By comparing the different combinations of the three modalities, it is found that the hierarchical fusion strategy that processes speech and text in advance can obviously improve the accuracy of the model. Experimental results on the public emotion datasets CH-SIMS and CMU-MOSI show that the proposed model achieves higher recognition accuracy than the baseline model, with three-class and two-class accuracy reaching 97.82% and 98.18% respectively, which proves the effectiveness of the model.
摘要:Attribute reduction, as a very important branch of formal concept analysis, is equally important in the three-way concept analysis. Based on object-induced concept lattices, attribute reduction theory which preserve OE-consistency of formal decision contexts is proposed, which enriches the reduction theory of three-way concept analysis. Firstly, OE-consistent set and OE-reduct of formal decision contexts are defined, and the attributes are classified into three categories according to their characteristics. Then, it is pointed out the essence of OE-reduct is minimal OE-consistent set, and several judgement theorems of OE-consistent set are given. By studying the necessary and sufficient conditions of OE-consistent set, judgement theorem of OE-reduct is obtained. Finally, the definition of OE-discernibility matrix and OE-discernibility function are given, and the method of calculating OE-reduct by using OE-discernibility matrix and OE-discernibility function is given.
摘要:As one of the important tools for knowledge discovery and data mining, rough set theory based on granular-ball computing has been successfully applied to label prediction and attribute reduction. However, the existing granular-ball rough set models only consider a single granulation, and cannot analyze and process data from a multi-granulation, and there are still many application scenarios that need to be considered from the perspective of multi-granulation. Based on this, this paper proposes a multi-granulation rough set based on granular-ball computing by embedding the idea of granular-ball in the multi-granulation rough set model, and discusses the relevant properties of the model. The model divides the data by setting the purity, which can effectively depict the internal relationship between the data, and thus design a position region generation algorithm for multi-granulation granular-ball rough set. Experimental analysis shows the feasibility and effectiveness of this model.
摘要:Concepts are the foundation for knowledge acquisition through formal concept analysis. In incomplete formal contexts, in order to express "jointly must possessing (possessed)" and "jointly might possessing (possessed)" relationships, the extent or intent of concepts is usually expressed in the form of interval set. We refer to such concepts as partially-known formal concepts. From the definition of partially-known formal concepts, it can be seen that their essence is related to the least and greatest completions of incomplete formal contexts. Therefore, we consider performing operations on the least and greatest completions of incomplete formal contexts to find partially-known formal concepts. We construct two new formal contexts based on the apposition and subposition of the least and greatest completions. Their concept lattices are isomorphic to the SE-ISI concept lattice and the ISE-SI concept lattice, respectively. Therefore, the new methods for constructing SE-ISI concept lattice and ISE-SI concept lattice are proposed. For the ISE-ISI concept, we use direct sum operation of the least and greatest completions to construct a new formal context, and propose a method to search for the ISE-ISI concepts.
关键词:incomplete formal context;partially-known formal concept;apposition;subposition;direct sum
摘要:Due to the limited distribution of lithium resources, the high cost disadvantage of lithium-ion battery gradually affects its large-scale application. In contrast, sodium resources are abundant in the world, as a substitute for lithium gradually gained widespread attention. However, due to the large radius of sodium ions, the material will undergo a large volume change during electrode insertion/removal, resulting in structural collapse. Therefore, in order to realize the high performance application of sodium electricity, it is necessary to develop cathode materials with long cycle life and good magnification performance. A novel polyanionic compound, niobium phosphate oxide Nb(PO4)O, was synthesized by one-step calcination with phytate-melamine supramolecular polymer assisted strategy. First of all, the material has a stable large frame structure to ensure the rapid transmission of sodium ions. Secondly, the loose sheet structure of the surface is conducive to the mass transfer between the electrolyte and the electrode, effectively shortening the diffusion path. When used as a sodium anode, the group of sodium ion batteries, in 1 000 mA·g-1 current density, 3 000 cycles after reversible discharge capacity can reach 133.7 mA·h·g-1, in 4 000 mA·g-1 high current density, reversible specific capacity is still 117.9 mA·h·g-1. Compared with Nb2O5, it has obvious magnification performance and long cycle stability.
关键词:sodium ion battery;negative electrode materials;polyanion compound;niobium phosphate oxide
摘要:Industrial wastewater treatment is of great significance for maintaining ecological balance and achieving sustainable development. Developing new removal agents with the feature of catalytic degradation of pollutants is an effective solution to the wastewater treatment. To address this, metal-organic frameworks (MOFs) with inherent characteristics of wide material sources and adjustable structures were proposed and demonstrated according to the principle of crystal engineering. Therefore, a MOF of [Cd(bpt)(H2O)4] n·n(pam)·2nH2O (MOF-1) was prepared by using pamoic acid(H2pm) and 2,5-bis (4-pyridyl)-1,3,4-thiadiazole (bpt) as components.The structural analysis and physico-chemical characterization showed that MOF-1possessed complex supramolecular network with one-dimensional covalent chain, the simple synthesis method, good product quality, and the uniform morphology with the size of about 0.5~ 2 μm. The main framework of MOF-1 was thermostable till 310 ℃. The crystal products owned excellent chemical stability of not less than 60 days and 163 days in strong acid and neutral aqueous solution, respectively. Moreover, MOF-1 displayed bright green fluorescence and could be used as a molecular luminescence thermometer over a wide temperature range.The catalytic degradation of organic dyes performed by MOF-1 in simulated industrial wastewater revealed that the materials had a narrow bandgap width Eg=2.85 eV and a large decolorization rate of Rhodamine B with 85.94%. Hence, MOF-1 is a brilliant wastewater treatment agent with greater comparative advantages with lots of known materials.
摘要:The rapid development of camellia oleifera industry has led to a large number of camellia oleifera seed residue, which caused serious waste of resources, restricting the sustainable development of China’s camellia oleifera industry.Considering the high contents of organic matter, using camellia oleifera seed residue to prepare organic fertilizer is an effective way for resource utilization. In the process of composting, cellulose, lignin and tea saponin and other biological macromolecules should be effectively biodegraded by microorganisms. In this paper, three highly efficient degrading strains targeting cellulose and lignin in camellia oleifera seed residue were obtained by screening and domestication, including 1 strain of fungus X1-1(Pleurostomarichardsiae), and 2 strains of bacteria M4-2 (Pseudomonas monteilii) and M5-2 (Paenibacillus sp.). The three strains can also effectively reduce the content of tea saponin in fermentation products, to realize the harmless treatment of camellia oleifera seed residue.The experimental results showed that after 28 days of fermentation with strain X1-1, M4-2 and M5-2, the degradation rates of cellulose content in camellia oleifera seed residue were 29.74%, 19.65% and 13.19%, the degradation rates of lignin content were 12.84%, 17.85% and 26.39%, and the degradation rates of tea saponin content were 34.06%, 45.05% and 46.15%, respectively.Moreover, the maturity, organic matter content, andnutrient content in fermentation product of camellia oleifera seed residue all met the organic fertilizer standards.The results of this study expand the resource utilization path of camellia oleifera seed residue, and provide good strain resources for the efficient degradation of lignin and cellulose.
摘要:Revealing the spatiotemporal characteristics and influencing factors of food supply service flow is a necessary prerequisite and important basis for ensuring food security and sustainable supply of this service. Taking Apple’s supply services as an example, a research theoretical framework was first constructed. Then, methods such as spatial autocorrelation and geographic detectors were used to analyze the spatiotemporal characteristics and influencing factors of Apple Provision Service Flow (APEF) under the influence of three main entities: merchants, enterprises, and e-commerce. The results show that ① the APEF time shows a seasonal variation, with the highest flow in autumn and the lowest in summer. The " peak" trend is significantly affected by holidays, and there are significant differences among different subjects. ②In terms of space, it is distributed in a " trumpet" shape with multiple cores, mainly flowing towards the eastern, southern, and southwestern regions of China. Among them, customers are in a radial shape with a " south-north" direction, enterprises are in a " three core and three line" shape, and e-commerce is in a " multi-core bead string" shape. ③The main influencing factors of APEF include apple production, economic development, and consumption level. The research conclusion can provide theoretical reference for solving the spatiotemporal mismatch between supply and demand of ecosystem services and scientific management.
关键词:ecosystem service flow;entities;spatiotemporal variations;influencing factors
摘要:Exploring the relationship between the livelihood capital of farmer and well-being is significant for improving the well-being of farmer in ecologically fragile areas and effectively reducing poverty. This study takes Dujiashigou of Mizhi County, Shaanxi Province as an example, which is located in the loess hilly and gully region. Through the constructed index system, the status quo of farmers’ livelihood capital and well-being were analyzed, and the Structural Equation Model was used to explore the impact of farmers’ livelihood capital on their well-being and its group differences. The results show that: ① the overall livelihood capital of farmer shows a spatial layout of " low value dispersed, high value concentrated", and there are significant spatial differences in the livelihood capital of different farmer groups. ② The spatial distribution pattern of objective well-being of farmers as a whole is " scattered at low values and locally clustered at high values", and the spatial distribution pattern of subjective well-being is " high around and low in the middle"; there are significant differences in the spatial distribution of subjective and objective well-being levels among farmer groups. ③ Farmers’ overall livelihood capital has a significant positive effect on well-being and objective well-being on subjective well-being, and there are significant differences in the effects of each sub-dimension of livelihood capital on farmers’ well-being. ④ There are group differences in the effects of farm households’ livelihood capital on well-being, pure agricultural farmers improved well-being through housing status, crop type and actual cultivated area per capita; crop type, housing status and linkage costs contributed more to part-time farmers; while household fixed assets, male labor force ratio and average age of household head mainly affected non-agricultural farmers. The study offers promising methods and approaches for improving the well-being of specific farmer groups.
关键词:farmers’ livelihood capital;objective well-being;subjective well-being;structural equation model;Mizhi County
摘要:The rural areas in urban areas have natural advantages in developing tourism, and how to maintain rural characteristics while developing rural tourism is an important practical and academic issue that needs attention at present. The study takes typical rural areas in the Xi'an metropolitan area as the case object, and based on the " starting point-driving force" analysis framework, constructs an adaptability evaluation model for rural tourism in urban areas, exploring different models and improvement paths for the development of rural tourism in urban areas. The results show that: ①The adaptability of rural tourism development in the Xi'an metropolitan area is at an average level; ②Based on the evaluation results, it can be divided into two modes: Adaptation in Development and Development in Adaptation. The former mainly focuses on the transformation and integration of rural industries driven by endogenous factors to develop rural tourism; the latter is implanted in rural tourism under the influence of external forces for rural spatial reconstruction. ③The study proposes to attach importance to the " starting point" role in the development of rural tourism in urban areas, promote the mutual adaptation of rural tourism, and strengthen the role of " driving force" to ensure the sustainability of rural tourism, in order to provide theoretical reference for the healthy development of rural tourism in China’s urban areas.
关键词:rural tourism;adaptability;"starting point-motive force" analytical framework;development model;Xi'an metropolitan area
摘要:Traditional villages are an important carrier for promoting rural revitalization and new urbanization construction. The spatial distribution characteristics of 182 national-level traditional villages in Shaanxi Province were analyzed as research objects. Using ArcGIS network analysis and GWR model, the spatial differentiation of traditional village accessibility characteristics and influencing factors at the county and town scales were analyzed in depth. The results of the study show that: ① The national-level traditional villages in Shaanxi Province are characterized by agglomeration spatial distribution, forming four main agglomeration area in space. ② The overall level of accessibility is low, showing a stepwise decline centered on Guanzhong. ③ There is strong spatial instability in the accessibility and influencing factors of traditional village accessibility at different spatial scales. At the county level, the elevation has a positive correlation with the accessibility of traditional villages. However, at the town level, both positive and negative relationships coexist, and are characterized by a gradual increase from the northwestern part of Yulin and the central part of Weinan to the peripheral circles. At the county level, the distance to neighboring rivers is negatively correlated with the level of accessibility of traditional villages. However, at the town level, its impact on accessibility shows a decreasing trend from southern and northern Shaanxi to the central part of the country. ④ Elevation, slope, and the distance to neighboring rivers are the main factors in the accessibility of national-level traditional villages in Shaanxi Province and have different influencing effects in different areas. Research on the accessibility of traditional villages can help reveal the ways to improve the accessibility of traditional villages in different regions, and can provide a scientific basis for the formulation of future conservation and utilization strategies for traditional villages.
关键词:traditional villages;accessibility;geographically weighted regression model;different scales;Shaanxi Province
摘要:With the rapid economic and social development of China, many countries around the world have ushered in a boom in the construction of overseas Chinese studies institutions, which is worth analyzing from the perspective of geography. Based on 378 overseas Chinese studies institutions data, the organizational structure, spatio-temporal pattern and driving mechanisms of overseas Chinese studies institutions are explored. This study finds that: ① Overseas Chinese studies institutions mainly rely on the construction of universities, and are mostly set up as independent institutions or divided into larger research areas such as East Asia, Asia, and the Asia-Pacific region, presenting a trend of simultaneous research in humanities and social sciences. ② The number of overseas Chinese studies institutions shows a " toothbrush-shaped" growth, which can be divided into three stages: initial exploration, vigorous development, adjustment and optimization, and the purpose in each stage is different, focusing on historical China to contemporary China. ③ The global spatial pattern of overseas Chinese studies institutions is extremely unbalanced, showing the pattern of " pan-north, dual-core, and multi-point", mainly driven by the internationalization level of higher education and economic and trade factors.Seeking economic and trade cooperation or drawing on economic development experience has become the third main motivation for the establishment of overseas Chinese studies institutions in addition to political and academic purposes. Finally, the problems existing in this study and the future directions are summarized and prospected.
关键词:overseas Chinese studies;research institutions;spatio-temporal pattern;driving factors;geodetector
摘要:This work focuses on the strike-slip faulting activity and stress field evolution in Honghe Oilfield, Ordos Basin. With the finite difference software, a series of 3D numerical models were constructed to investigate the two stages of structural superposition. Constrained by the refined seismic structure interpretation and the tectonic evolution of Qinling Mountain in the south and Liupanshan Mountain in the west, the model results indicate that the NW trending Yudu Fault developed first and was controlled by the sinistral strike-slip of basement fault. The Yudu Fault is supposed to be active in the late Jurassic-early Cretaceous, which is mainly attributed to the northwestward compression among different blocks in the northern margin of the Qinling Mountains. The maximum principal stress near the Yudu Fault Zone was oriented NEE during faulting. Subsequently, the NEE trending faults developed and were subjected to distributed sinistral strike-slip faulting during eastward compression of the Liupanshan thrust belt. These faults developed in the late Early Cretaceous or the Cenozoic. The maximum principal stress was oriented in the range of N36°~69° with an average of N53°. This study provides a new constraint on the fault prediction in this area, and also sheds new light on faulting dynamics in the other oilfields in the Ordos Basin.
摘要:The study focuses on the burnt rocks of the 22 and 31 sections of the Yan’an Formation in Zhangjiamao Coalfield, Northern Shaanxi, and analyzes their geochemical migration patterns and metamorphic mineral facies characteristics. The results indicate that the burnt rocks in this area are low alkaline silicoaluminal metamorphic rocks, with uneven distribution of calcium and alkali horizontally. A small number of burnt rock samples have high loss on ignition and high CaO content, with Fe elements migrate and accumulate. Zr has better compatibility than Hf, and the enrichment of Ba element may be related to later hydrological processes. The trend line of rare earth element distribution is little L-REE enrichment, and the effect of burning on the distribution of rare earth elements is relatively limited. The primary minerals in burnt rocks often exhibit ablative characteristics, with the development of ablative edges and pore structures under the microscope. Newly formed minerals form aggregates and fill mineral spaces or fractures. The mineral facies combination is similar to the diorite facies that undergoes high thermal metamorphism during thermal contact metamorphism, but the metamorphic pressure is lower and the temperature is higher. The burnt rock layers can be divided from bottom to top into a combination of thin plate-like burnt rock, burnt lava, thick plate-like burnt rock, and thick layered burnt rock. tridymite+ cristobalite content can better characterize the burning degreecompared to mullite+ cordierite+ hematite content. The degree of burning in the lower part of the same layer burnt rock is higher than that in the upper part, and the burning degree of plate-like burnt rock in the upper part of the burnt lava is higher than that in the lower part. The burnt rocks in the research area have the " hidden" characteristic and have the same sedimentary period burnt effect. Except for the burnt rock layers exposed in the valley terrace, some of the burnt rock layers are buried under the Pleistocene strata and residual yan’an Formation strata. The early formation of burnt rock layers was later destroyed, and multiple stages of burnt rock layers overlapped to form the current burnt area.
关键词:burnt rock;Zhangjiamao minefield;geochemistry;metamorphic mineral facies;thermal contact metamorphism;diorite facies