Straight Forward Constructive Deep Learning Neural Network (SFC-DLNN) algorithm is a new architecturebased algorithm for artificial neural networks. Rather than simply adjusting the weights in a fixed topology network, SFC-DLNN starts with a minimal topology (perceptron), then builds up their network by gradually trains and adds new nodes one by one, creating multiple layers’ network. Once a unit has been added to the network, the weights of the new architecture are generated. This unit then stands as a permanent detector of features in the network, and a more complex feature space is then created where the data is likely to be linearly separable. The SFC-DLNN algorithm has many advantages over existing ones: it has good learning speed, the network determines its topology size, and the structures it has built is retained after the training stage. We obtain from our built model (SFC-DLNN) an accuracy and specificity of 83:5% from a simulated data set using the uniform distribution. This is not the best but is enough to approve the model prediction capacity
{"title":"Straight Forward Constructive Deep Learning Neural Network (SFC-DLNN) Algorithm","authors":"","doi":"10.33140/amlai.03.02.04","DOIUrl":"https://doi.org/10.33140/amlai.03.02.04","url":null,"abstract":"Straight Forward Constructive Deep Learning Neural Network (SFC-DLNN) algorithm is a new architecturebased algorithm for artificial neural networks. Rather than simply adjusting the weights in a fixed topology network, SFC-DLNN starts with a minimal topology (perceptron), then builds up their network by gradually trains and adds new nodes one by one, creating multiple layers’ network. Once a unit has been added to the network, the weights of the new architecture are generated. This unit then stands as a permanent detector of features in the network, and a more complex feature space is then created where the data is likely to be linearly separable. The SFC-DLNN algorithm has many advantages over existing ones: it has good learning speed, the network determines its topology size, and the structures it has built is retained after the training stage. We obtain from our built model (SFC-DLNN) an accuracy and specificity of 83:5% from a simulated data set using the uniform distribution. This is not the best but is enough to approve the model prediction capacity","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128222249","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Drones have piqued the interest of the mining industry, which has expressed a strong interest in the usage of UAVs/drones for regular tasks Unmanned Aerial Vehicles(UAVs)/drones, sometimes known as Micro Air Vehicles (MAVs), are mostly drones that are used for a number of commercial and military applications, including surveillance and reconnaissance. These unmanned aerial vehicles (UAVs)/drones are capable of transporting of a wide range sensors depending on the nature of their missions, including acoustic, optical, biochemical, and bio sensors. In order to improve the performance and efficiency of drones/UAV, researchers have concentrated on the design optimization of drones, which has resulted in the creation and construction of a variety of Aerial Vehicles/drones with diverse abilities and capabilities. As a consequence, previous research as well as information from firms that supply drones for the mining industry are being explored further. An investigation of the application of drone/UAVs in surface and subsurface mines is presented in this research. The usage of drones/UAVs in abandoned mines, both on the surface and below, is also discussed. It also includes a thorough discussion of the instruments or sensors that are frequently used in mining drones. In this paper/article, we address the difficulties linked with the usage of drones technologies in underground mines, as well as potential solutions to these difficulties.
{"title":"Unmanned Aerial Vehicle Application in Mining User case in Rwanda","authors":"","doi":"10.33140/amlai.03.02.05","DOIUrl":"https://doi.org/10.33140/amlai.03.02.05","url":null,"abstract":"Drones have piqued the interest of the mining industry, which has expressed a strong interest in the usage of UAVs/drones for regular tasks Unmanned Aerial Vehicles(UAVs)/drones, sometimes known as Micro Air Vehicles (MAVs), are mostly drones that are used for a number of commercial and military applications, including surveillance and reconnaissance. These unmanned aerial vehicles (UAVs)/drones are capable of transporting of a wide range sensors depending on the nature of their missions, including acoustic, optical, biochemical, and bio sensors. In order to improve the performance and efficiency of drones/UAV, researchers have concentrated on the design optimization of drones, which has resulted in the creation and construction of a variety of Aerial Vehicles/drones with diverse abilities and capabilities. As a consequence, previous research as well as information from firms that supply drones for the mining industry are being explored further. An investigation of the application of drone/UAVs in surface and subsurface mines is presented in this research. The usage of drones/UAVs in abandoned mines, both on the surface and below, is also discussed. It also includes a thorough discussion of the instruments or sensors that are frequently used in mining drones. In this paper/article, we address the difficulties linked with the usage of drones technologies in underground mines, as well as potential solutions to these difficulties.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131312380","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The COVID-19 (coronavirus) is an ongoing pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The virus was first identified in mid-December 2019 in the Hubei province of Wuhan, China and by now has spread throughout the planet with more than 75.5 million confirmed cases and more than 1.67 million deaths. With limited number of COVID-19 test kits available in medical facilities, it is important to develop and implement an automatic detection system as an alternative diagnosis option for COVID-19 detection that can used on a commercial scale. Chest X-ray is the first imaging technique that plays an important role in the diagnosis of COVID-19 disease. Computer vision and deep learning techniques can help in determining COVID-19 virus with Chest X-ray Images. Due to the high availability of large-scale annotated image datasets, great success has been achieved using convolutional neural network for image analysis and classification. In this research, we have proposed a deep convolutional neural network trained on five open access datasets with binary output: Normal and Covid. The performance of the model is compared with four pre-trained convolutional neural network- based models (COVID-Net, ResNet18, ResNet and MobileNet-V2) and it has been seen that the proposed model provides better accuracy on the validation set as compared to the other four pre-trained models. This research work provides promising results which can be further improvise and implement on a commercial scale.
{"title":"A Deep Learning Approach for the Detection of COVID-19 from Chest X-Ray images using Convolutional Neural Networks","authors":"Aditya Singh Shamsheer Pal Saxena","doi":"10.33140/amlai.03.02.01","DOIUrl":"https://doi.org/10.33140/amlai.03.02.01","url":null,"abstract":"The COVID-19 (coronavirus) is an ongoing pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The virus was first identified in mid-December 2019 in the Hubei province of Wuhan, China and by now has spread throughout the planet with more than 75.5 million confirmed cases and more than 1.67 million deaths. With limited number of COVID-19 test kits available in medical facilities, it is important to develop and implement an automatic detection system as an alternative diagnosis option for COVID-19 detection that can used on a commercial scale. Chest X-ray is the first imaging technique that plays an important role in the diagnosis of COVID-19 disease. Computer vision and deep learning techniques can help in determining COVID-19 virus with Chest X-ray Images. Due to the high availability of large-scale annotated image datasets, great success has been achieved using convolutional neural network for image analysis and classification. In this research, we have proposed a deep convolutional neural network trained on five open access datasets with binary output: Normal and Covid. The performance of the model is compared with four pre-trained convolutional neural network- based models (COVID-Net, ResNet18, ResNet and MobileNet-V2) and it has been seen that the proposed model provides better accuracy on the validation set as compared to the other four pre-trained models. This research work provides promising results which can be further improvise and implement on a commercial scale.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-04-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134127789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper elaborates the recent enhancements on IoT capabilities and efficiencies. From power, coverage, cost, complexity, device density, core network protocol, spectrum efficiency perspectives, it describes a comprehensive blueprint for driving IoT optimizations. For better mobile broadband experience, enabling Gigabit-class throughput with advanced 5G network techniques, millimeter wave, massive MIMO, carrier aggregation, and LAA benefit massive IoT improvements. Supporting Gigabit-class data rates for high-performance IoT requires high power efficiency. eMTC (enhanced machine-type communication) optimizes for the broadest range of IoT applications with VoLTE and mobility. NB-IoT (narrowband IoT) provides optimizations for high throughput and low delay LPWAN IoT use cases. Index Terms-LTE IoT, eMTC, NB-IoT, QoS, QoE.
{"title":"Efficient Resource Optimization of 5G Networks Enabling QoS and QoE in IoT Applications","authors":"","doi":"10.33140/amlai.03.01.05","DOIUrl":"https://doi.org/10.33140/amlai.03.01.05","url":null,"abstract":"This paper elaborates the recent enhancements on IoT capabilities and efficiencies. From power, coverage, cost, complexity, device density, core network protocol, spectrum efficiency perspectives, it describes a comprehensive blueprint for driving IoT optimizations. For better mobile broadband experience, enabling Gigabit-class throughput with advanced 5G network techniques, millimeter wave, massive MIMO, carrier aggregation, and LAA benefit massive IoT improvements. Supporting Gigabit-class data rates for high-performance IoT requires high power efficiency. eMTC (enhanced machine-type communication) optimizes for the broadest range of IoT applications with VoLTE and mobility. NB-IoT (narrowband IoT) provides optimizations for high throughput and low delay LPWAN IoT use cases. Index Terms-LTE IoT, eMTC, NB-IoT, QoS, QoE.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133248513","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper proposes a prioritization-based distributed storage of big data processing application in Internet of Vehicle (IoV) system. Designing a scalable, high-performance big data distributed storage system for IoV, an advanced data-processing system for car services. The novel contribution focused on developing vehicular multi-channel control protocol that control the prioritization of services, according to bit rate, transmit power, speed, inter-vehicle distance. The proposed scheme can achieve higher performance in IoV storage system. Index Terms- IoV, sensor fusion, distributed storage, prioritization, edge computing, cloud computing.
{"title":"Blockchain Towards Prioritization-Based Distributed Storage of Big Data for Internet of Vehicles","authors":"","doi":"10.33140/amlai.03.01.04","DOIUrl":"https://doi.org/10.33140/amlai.03.01.04","url":null,"abstract":"This paper proposes a prioritization-based distributed storage of big data processing application in Internet of Vehicle (IoV) system. Designing a scalable, high-performance big data distributed storage system for IoV, an advanced data-processing system for car services. The novel contribution focused on developing vehicular multi-channel control protocol that control the prioritization of services, according to bit rate, transmit power, speed, inter-vehicle distance. The proposed scheme can achieve higher performance in IoV storage system. Index Terms- IoV, sensor fusion, distributed storage, prioritization, edge computing, cloud computing.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122635381","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Roberts-Janet Nuclear Periodic Table has emerged recently. The inversion of the Periodic Table to accommodate spatial variation of atomic energy levels relative to the nucleus has subsequently been underwritten by Quantum Field Theory’s U (1) X SU (2) x SU (3) group symmetry and Clifford Algebra resulting in a one-toone mapping between the Roberts-Janet Table and The Quantum Mechanical Table. This manuscript attempts to show the over-arching nature of the Roberts-Janet Table epitomised by two cycles. The first of these is the role of causality within the lower half of the table in nucleosynthesis and cosmology whilst the second attempts to outline causality’s path in the upper half of the table in biochemical settings. The link between the cycles is the set of elements themselves; within theoretically an infinite group of elements as radioactivity is reignited having been extinguished temporarily in the ebb and flow of production and annihilation of white dwarfs, neutron stars and black holes. The current scientific landscape is outlined to create a platform from which to proceed. Various sizes of black hole production suggest a hierarchy of outcomes which produces a reignition of radioactivity and potentially a creation of other universes from the explosions of larger supermassive black holes as energies increase to the Planck scale resulting in periods of inflation and condensation that predate quark production. Universes could be superimposed on previous universes explaining why some supermassive black holes appear nearer than current theoretical models.
最近出现了罗伯特-珍妮特核元素周期表。元素周期表的反转，以适应相对于原子核的原子能级的空间变化，随后被量子场论的U (1) X SU (2) X SU(3)群对称和Clifford代数所支持，导致罗伯茨-珍妮特表和量子力学表之间的一对一映射。这篇手稿试图展示由两个周期体现的罗伯特-珍妮特表的总体性质。其中第一个是因果关系在核合成和宇宙学中表的下半部分的作用，而第二个试图概述因果关系在生化设置表的上半部分的路径。循环之间的联系是一组元素本身;理论上，在白矮星、中子星和黑洞的产生和湮灭的涨落过程中，放射性被暂时熄灭后，无限的一组元素被重新点燃。概述当前的科学景观，以创建一个平台，从中进行。不同大小的黑洞的产生表明了一个层次的结果，它产生了放射性的重新点燃，并可能从更大的超大质量黑洞的爆炸中创造出其他宇宙，因为能量增加到普朗克尺度，导致了夸克产生之前的膨胀和冷凝时期。宇宙可能叠加在先前的宇宙上，这解释了为什么一些超大质量黑洞看起来比目前的理论模型更近。
{"title":"Letter To Feynman, Einstein, Wallace, Darwin, Maxwell and Mendeleev","authors":"","doi":"10.33140/amlai.03.01.07","DOIUrl":"https://doi.org/10.33140/amlai.03.01.07","url":null,"abstract":"The Roberts-Janet Nuclear Periodic Table has emerged recently. The inversion of the Periodic Table to accommodate spatial variation of atomic energy levels relative to the nucleus has subsequently been underwritten by Quantum Field Theory’s U (1) X SU (2) x SU (3) group symmetry and Clifford Algebra resulting in a one-toone mapping between the Roberts-Janet Table and The Quantum Mechanical Table. This manuscript attempts to show the over-arching nature of the Roberts-Janet Table epitomised by two cycles. The first of these is the role of causality within the lower half of the table in nucleosynthesis and cosmology whilst the second attempts to outline causality’s path in the upper half of the table in biochemical settings. The link between the cycles is the set of elements themselves; within theoretically an infinite group of elements as radioactivity is reignited having been extinguished temporarily in the ebb and flow of production and annihilation of white dwarfs, neutron stars and black holes. The current scientific landscape is outlined to create a platform from which to proceed. Various sizes of black hole production suggest a hierarchy of outcomes which produces a reignition of radioactivity and potentially a creation of other universes from the explosions of larger supermassive black holes as energies increase to the Planck scale resulting in periods of inflation and condensation that predate quark production. Universes could be superimposed on previous universes explaining why some supermassive black holes appear nearer than current theoretical models.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122610269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Objective: There are numerous functional magnetic resonance imaging (fMRI) studies examining the cerebral function of migraine patients using regional homogeneity (ReHo) and functional connectivity (FC) measurements. However, these studies generally report inconsistent conclusions. We will performed a systematic review and meta-analysis of this body of literature, aiming to identify consistent conclusions regarding cerebral functional changes in migraine patients and to describe potential future directions.Methods: Two investigators will independently screen studies published in online databases (i.e., Medline, Cochrane Library, PubMed, and Web of Science) from the database inception to June 1, 2021. By discussing with a third investigator, any disagreement will be resolved and will attain consensus. A coordinate-based meta-analysis will then be performed with an activation likelihood estimate (ALE) random-effects model.Results: The cerebral FC and ReHo altered regions in migraine patients will be elucidated in this meta-analysis.Conclusion: This study will reveal cerebral functional changes of migraine patients based on current literature to identify consistent conclusions and to describe potential future direction.Registration number: CRD42021257300.
目的:有许多功能性磁共振成像(fMRI)研究使用区域均匀性(ReHo)和功能连通性(FC)测量来检查偏头痛患者的大脑功能。然而，这些研究通常报告不一致的结论。我们将对这些文献进行系统回顾和荟萃分析，旨在确定偏头痛患者脑功能改变的一致结论，并描述潜在的未来发展方向。方法:两名研究者将独立筛选从数据库建立到2021年6月1日在在线数据库(即Medline、Cochrane Library、PubMed和Web of Science)中发表的研究。通过与第三方调查员的讨论，任何分歧都将得到解决并达成共识。然后使用激活似然估计(ALE)随机效应模型进行基于坐标的元分析。结果:本荟萃分析将阐明偏头痛患者的大脑FC和ReHo改变区域。结论:本研究将在现有文献的基础上揭示偏头痛患者的脑功能变化，以确定一致的结论并描述潜在的未来方向。注册号:CRD42021257300。
{"title":"Functional Connectivity And Regional Homogeneity Alterations In Migraine Patients: A Protocol of Systematic Review And Meta-Analysis","authors":"Yuzhong Cui, Q. Xu, Yu-Ting Li, Yanguo Zhang, Jing-Ting Sun, Ze-Yang Li, Min-Hua Ni, Teng Ma, Lin-Feng Yan, G. Cui, Wen Wang, Zhuanghong Chen","doi":"10.21203/rs.3.rs-1150880/v1","DOIUrl":"https://doi.org/10.21203/rs.3.rs-1150880/v1","url":null,"abstract":"\u0000 Objective: There are numerous functional magnetic resonance imaging (fMRI) studies examining the cerebral function of migraine patients using regional homogeneity (ReHo) and functional connectivity (FC) measurements. However, these studies generally report inconsistent conclusions. We will performed a systematic review and meta-analysis of this body of literature, aiming to identify consistent conclusions regarding cerebral functional changes in migraine patients and to describe potential future directions.Methods: Two investigators will independently screen studies published in online databases (i.e., Medline, Cochrane Library, PubMed, and Web of Science) from the database inception to June 1, 2021. By discussing with a third investigator, any disagreement will be resolved and will attain consensus. A coordinate-based meta-analysis will then be performed with an activation likelihood estimate (ALE) random-effects model.Results: The cerebral FC and ReHo altered regions in migraine patients will be elucidated in this meta-analysis.Conclusion: This study will reveal cerebral functional changes of migraine patients based on current literature to identify consistent conclusions and to describe potential future direction.Registration number: CRD42021257300.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128647721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-06-23DOI: 10.20944/preprints202106.0573.v1
Tekalign Regasa Ashale
In this paper, improved matrix Reduction Method is proposed for the solution of fuzzy transportation problem in which all inputs are taken as fuzzy numbers. Since ranking fuzzy number is important tool in decision making, Fuzzy trapezoidal number is converting in to crisp set by using Mean techniques and solved by proposed method for fuzzy transportation problem. We give suitable numerical example for unbalanced and compare the optimal value with other techniques. The Result shows that the optimum profit of transportation problem using proposed technique under robust ranking method is better than the other method. Novelty: The numerical illustration demonstrates that the new projected method for managing the transportation problems on fuzzy algorithms.
{"title":"Newly Proposed Matrix Reduction technique Under Mean Ranking Method for Solving Trapezoidal Fuzzy Transportation problems Under Fuzzy Environment","authors":"Tekalign Regasa Ashale","doi":"10.20944/preprints202106.0573.v1","DOIUrl":"https://doi.org/10.20944/preprints202106.0573.v1","url":null,"abstract":"In this paper, improved matrix Reduction Method is proposed for the solution of fuzzy transportation problem in which all inputs are taken as fuzzy numbers. Since ranking fuzzy number is important tool in decision making, Fuzzy trapezoidal number is converting in to crisp set by using Mean techniques and solved by proposed method for fuzzy transportation problem. We give suitable numerical example for unbalanced and compare the optimal value with other techniques. The Result shows that the optimum profit of transportation problem using proposed technique under robust ranking method is better than the other method. Novelty: The numerical illustration demonstrates that the new projected method for managing the transportation problems on fuzzy algorithms.","PeriodicalId":186756,"journal":{"name":"Advances in Machine Learning & Artificial Intelligence","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125559195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：