检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Lin, Xiaoqiang Xu, Xinyi Wu, Zhaoxuan Ng, See-Kiong Low, Bryan Kian Hsiang Department of Computer Science National University of Singapore Singapore Institute of Data Science National University of Singapore Singapore

data valuation quantifies the contribution of each data point to the performance of a machine learning model. Existing works typically define the value of data by its improvement of the validation performance of the trained model. However, this approach can be impractical to apply in collaborative machine learning and data marketplace since it is difficult for the parties/buyers to agree on a common validation dataset or determine the exact validation distribution a priori. To address this, we propose a distributionally robust data valuation approach to perform data valuation without known/fixed validation distributions. Our approach defines the value of data by its improvement of the distributionally robust generalization error (DRGE), thus providing a worst-case performance guarantee without a known/fixed validation distribution. However, since computing DRGE directly is infeasible, we propose using model deviation as a proxy for the marginal improvement of DRGE (for kernel regression and neural networks) to compute data values. Furthermore, we identify a notion of uniqueness where low uniqueness characterizes low-value data. We empirically demonstrate that our approach outperforms existing data valuation approaches in data selection and data removal tasks on real-world datasets (e.g., housing price prediction, diabetes hospitalization prediction). Copyright 2024 by the author(s)

关键词：

来源：评论

学校读者我要写书评

暂无评论

Community Clustering on Fraud Transactions Applied the Louvain-Coloring Algorithm

引用

International Journal of Electronics and Telecommunications 2023年第3期69卷 593-598页

作者： Mardiansyah, Heru Suwilo, Saib Nababan, Erna Budiarti Efendi, Syahril Department Computer Science Faculty of Computer Science and Information Technology Universitas Sumatera Utara Medan Indonesia Department Mathematics Faculty of Mathematics and Natural Sciences Universitas Sumatera Utara Medan Indonesia Department Data Science and Artificial Intelligence Faculty of Computer Science and Information Technology Universitas Sumatera Utara Medan Indonesia

The contribution main from this research is modularity and better processing time in detecting community by using K-1 coloring. Testing performed on transaction datasets remittance on P2P platforms where the Louvain Coloring algorithm is better in comparison to Louvain Algorithm data used is data transfer transactions made by customers on the P2P Online platform. The data is the User data that has information transfer transactions, Card data that has information card, IP data that has IP information, and Device data that has information device. Every user owns unique 128-bit identification, and other nodes representing card, device, and IP are assigned a random UUID. The Device node has the guide, and device properties. IP nodes only have property guide and node User has property fraud Money Transfer, guide, money Transfer Error Cancel Amount, first Charge back Date. Each node has a unique 128-bit guide, with the amount whole of as many as 789,856 nodes. Application technique K-1 staining on Louvain algorithm shows enhancement value modularity and better processing time for detecting community on the network large scale. Through a series of exercises and tests carried out in various scenarios, it shows that the experiments carried out in this paper, namely the Louvain Coloring algorithm, are more effective and efficient than the Louvain algorithm in scenario 1,3, and 5 meanwhile For Scenarios 2 and 4 Louvain Algorithm is better. © 2023 Polish Academy of sciences. All rights reserved.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Anomaly Detection in Network Traffic data Using Deep Learning 2

Anomaly Detection in Network Traffic Data Using Deep Learnin...

引用

2nd International Conference on Computing and data Analytics, ICCDA 2024

作者： Ramya, S. Selvi, S.Annal Ezhil Bishop Heber Collge Department of Data Science Thiruchirappalli India Bishop Heber College Department of Computer Science Thiruchirappalli India

ISBN: (纸本)9798331542573

In the Tech Renaissance, spotting network traffic anomalies has become a game changer for security. With the rapid growth of network traffic and the increasing frequency of cyberattacks, detecting anomalies and intrusions in real-time is critical. The growing complexity and volume of modern traffic patterns have rendered traditional security measures inadequate for identifying sophisticated cyber threats. This research aims to detect unusual behaviors in network traffic that could signal security breaches, misuse, or attacks. Current detection methods, like rule-based intrusion detection systems (IDS) and signature-based approaches, depend on known threat signatures and predefined rules. These methods are limited in detecting new or unknown threats, prone to high false positive rates, and lack the adaptability needed to respond to evolving threats. To address these limitations, this paper proposes the use of advanced machine learning and deep learning techniques, such as autoencoders, RNNs, CNNs, and GANs, which can learn from evolving traffic patterns and improve detection accuracy. By adapting to both known and unknown anomalies, these models significantly reduce false positives and enhance the overall security of network infrastructures. Organizations and businesses will benefit from improved detection rates, a reduced risk of cyberattacks, and enhanced protection of sensitive data, leading to a more secure and resilient network environment. The proposed models anticipate a notable reduction in false positives and an increase in anomaly detection accuracy, making them a more reliable solution than traditional methods. © 2024 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

Comparison of the SOLA model and ARIMA for Cloud Computing Resources Allocations 4

Comparison of the SOLA model and ARIMA for Cloud Computing R...

引用

4th International Multidisciplinary Information Technology and Engineering Conference, IMITEC 2024

作者： Prince Sekwatlakwatla, Sello Malele, Vusumuzi North-West University Unit for Data Science and Computing Department of Computer Science and Information Systems South Africa

ISBN: (纸本)9798350387988

Cloud computing solutions are becoming more and more popular as a way for organizations to improve productivity, save costs, and simplify procedures. The advantage of cloud services is that they enable users to store data remotely and access on-demand applications and services from a shared pool of configurable computing resources. Cloud computing resource allocation faces challenges in load balancing, effective resource management, and compliance with regulatory and legal requirements, requiring complex tools and techniques for optimal use. Most organizations are moving to cloud services and encouraging their clients to access services online. However, effective resource allocation is critical for improving performance and lowering costs in this area. Due to unpredictable network traffic in cloud computing, resource allocation is challenging, which causes customers to complain about application timeouts, delayed system times, and higher bandwidth use during peak hours. This entails allocating resources to various users and programs, such as memory, processing power, storage, and network bandwidth. In this regard, this study compares the ensemble method, which is stepwise Gaussian Linear Autoregressive (SGLA) and the individual method, which is autoregressive integrated moving average (ARIMA). The results show SGLA prediction accuracy increased with an average of 97.9%, and ARIMA prediction showed an accuracy of 71.5%. In this regard the ensemble method performed better than individual methods using the same datasets. The study recommends the ensemble method for the prediction and allocation of resources in cloud computing. © 2024 IEEE.

关键词： Cloud platforms

来源：评论

学校读者我要写书评

暂无评论

Plant Fettle Detector: Unifying Fruit Quality Tracking and Plant Disease Detection through Advanced Image Analysis 3

Plant Fettle Detector: Unifying Fruit Quality Tracking and P...

引用

3rd International Conference for Innovation in Technology, INOCON 2024

作者： Nair, Sruthi Bajpai, Khushi Fadnavis, Sakshi Jain, Shruti Sheikh, Zeenat Shri Ramdeobaba College of Engineering and Management Department of Computer Science and Engineering - Data Science Nagpur44013 India

ISBN: (纸本)9798350381931

India being an agricultural country, food quality tracking is a major challenge faced by common farmers across the country. This research presents an innovative integration of Convolutional Neural Networks (CNNs) to address key challenges in agriculture - fruit quality tracking and plant disease detection. The unified system comprises the Fruit Quality Tracker and the Plant Disease Detector, leveraging advanced image analysis for precision agriculture. The Fruit Quality Tracker uses computer vision to assess ripeness and defects in real-time, empowering farmers with informed decisions. Simultaneously, the Plant Disease Detector swiftly identifies and classifies diseases, mitigating economic losses. The research explores architectural nuances, development methodologies, and deployment on Google Cloud Platform. Testing shows high accuracy rates of 98.5% for fruit quality and 98.83% for plant disease detection. The system's transformative potential in agriculture is highlighted, bridging the gap between technology and sustainable food production systems. © 2024 IEEE.

关键词： Image analysis

来源：评论

学校读者我要写书评

暂无评论

Towards Adaptation in Multiobjective Evolutionary Algorithms for Integer Problems 13

Towards Adaptation in Multiobjective Evolutionary Algorithms...

引用

13th IEEE Congress on Evolutionary Computation, CEC 2024

作者： Rudolph, Günter Wagner, Markus Tu Dortmund University Department of Computer Science Dortmund Germany Monash University Department of Data Science and Ai Clayton Australia

ISBN: (纸本)9798350308365

Parameter control refers to the techniques that dynamically adapt the parameter values of the evolutionary algorithm during the optimization process, such as population size, crossover rate, or operator selection. Adaptation can improve the performance and robustness of the algorithm, however, parameter control mechanisms themselves need to be designed and configured carefully. With this article, we contribute a systematic investigation of an adaptive, multi-objective algorithm that is designed for the optimisation of problems in unbounded integer decision spaces. We find that (1) adaptation outperforms the best static configurations by 39-82 %, and (2) performance of the multi-objective algorithm is often independent of the adaptation scheme's initial configuration. © 2024 IEEE.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Iterative Methods via Locally Evolving Set Process 38

Iterative Methods via Locally Evolving Set Process

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Zhou, Baojian Sun, Yifan Harikandeh, Reza Babanezhad Guo, Xingzhi Yang, Deqing Xiao, Yanghua The School of Data Science Fudan University China Shanghai Key Laboratory of Data Science School of Computer Science Fudan University China Department of Computer Science Stony Brook University United States Samsung SAIT AI Lab

Given the damping factor α and precision tolerance ϵ, Andersen et al. [2] introduced Approximate Personalized PageRank (APPR), the de facto local method for approximating the PPR vector, with runtime bounded by Θ(1/(αϵ)) independent of the graph size. Recently, Fountoulakis & Yang [12] asked whether faster local algorithms could be developed using Õ(1/(√αϵ)) operations. By noticing that APPR is a local variant of Gauss-Seidel, this paper explores the question of whether standard iterative solvers can be effectively localized. We propose to use the locally evolving set process, a novel framework to characterize the algorithm locality, and demonstrate that many standard solvers can be effectively localized. Let vol(St) and γt be the running average of volume and the residual ratio of active nodes St during the process. We show vol(St)/γt ≤ 1/ϵ and prove APPR admits a new runtime bound Õ(vol(St)/(αγt)) mirroring the actual performance. Furthermore, when the geometric mean of residual reduction is Θ(√α), then there exists c ∈ (0, 2) such that the local Chebyshev method has runtime Õ(vol(St)/(√α(2 − c))) without the monotonicity assumption. Numerical results confirm the efficiency of this novel framework and show up to a hundredfold speedup over corresponding standard solvers on real-world graphs. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Artificial Neural Network-Based Model for Effective Software Development Effort Estimation

引用

computer Systems science & Engineering 2023年第2期44卷 1309-1324页

作者： Junaid Rashid Sumera Kanwal Muhammad Wasif Nisar Jungeun Kim Amir Hussain Department of Computer Science and Engineering Kongju National UniversityCheonan31080Korea Department of Computer Science COMSATS University IslamabadWah CampusIslamabad47040Pakistan Centre of AI and Data Science Edinburgh Napier UniversityEdinburghEH114DYUK

In project management,effective cost estimation is one of the most cru-cial activities to efﬁciently manage resources by predicting the required cost to fulﬁll a given ***,ﬁnding the best estimation results in software devel-opment is ***,accurate estimation of software development efforts is always a concern for many *** this paper,we proposed a novel soft-ware development effort estimation model based both on constructive cost model II(COCOMO II)and the artiﬁcial neural network(ANN).An artiﬁcial neural net-work enhances the COCOMO model,and the value of the baseline effort constant A is calibrated to use it in the proposed model *** state-of-the-art publicly available datasets are used for *** backpropagation feed-forward procedure used a training set by iteratively processing and training a neural *** proposed model is tested on the test *** estimated effort is compared with the actual effort *** results show that the effort estimated by the proposed model is very close to the real effort,thus enhanced the reliability and improving the software effort estimation accuracy.

关键词： Software cost estimation neural network backpropagation forward neural networks software effort estimation artiﬁcial neural network

来源：评论

学校读者我要写书评

暂无评论

Customer Segmentation Based on Recency, Frequency, Monetary, Variety and Duration (RFMVD) 2

Customer Segmentation Based on Recency, Frequency, Monetary,...

引用

2nd International Conference on Technology Innovation and Its Applications, ICTIIA 2024

作者： Sundari, Agus Lydia, Maya Silvi Muchtar, Muhammad Anggia Universitas Sumatera Utara Faculty of Computer Science and Information Technology Department of Data Science and Artificial Intelligence Medan Indonesia Universitas Sumatera Utara Faculty of Computer Science and Information Technology Department of Computer Science Medan Indonesia

ISBN: (纸本)9798350351613

In banking, maintaining customer retention and customer satisfaction are important. Effective customer segmentation can be a strategic tool to improve customer loyalty and business performance. This research can assist banks in developing marketing strategies to retain customers and improve services based on bill payment transactions. The data that will be used in this study is bill payment transaction data at regional banks in 2023, as for the types of bill payments made by customers, namely electricity, water, telephone payments, and internet bills. The amount of data used is 702,174 rows. This research uses K-Means clustering to group customers based on recency (the last time the customer made a transaction), frequency (the number of customers making transactions), monetary (the nominal amount of money spent by the customer), variety (the number of types of bills paid) and duration (the average duration of time the customer makes the previous transaction until the next transaction) (RFMVD). The results of customer segmentation based on recency, frequency, monetary, variety and duration using K-Means clustering produce 3 customer groups, namely passive customers, loyal customers, and VIP customers. The silhouette score value on customer segmentation using K-Means clustering is 0.7303, which indicates that the grouping is quite good and the number of clusters is correct. © 2024 IEEE.

关键词： bill payments Customer segmentation K-Means clustering RFMVD Model silhouette score

来源：评论

学校读者我要写书评

暂无评论

Dual VC Dimension Obstructs Sample Compression by Embeddings 37

Dual VC Dimension Obstructs Sample Compression by Embeddings

引用

37th Annual Conference on Learning Theory, COLT 2024

作者： Chase, Zachary Chornomaz, Bogdan Hanneke, Steve Moran, Shay Yehudayoff, Amir Department of Mathematics Technion Israel Department of Computer Science Purdue University United States Departments of Mathematics Computer Science and Data and Decision Sciences Technion Google Research Israel Department of Computer Science Copenhagen University Denmark

This work studies embedding of arbitrary VC classes in well-behaved VC classes, focusing particularly on extremal classes. Our main result expresses an impossibility: such embeddings necessarily require a significant increase in dimension. In particular, we prove that for every d there is a class with VC dimension d that cannot be embedded in any extremal class of VC dimension smaller than exponential in d. In addition to its independent interest, this result has an important implication in learning theory, as it reveals a fundamental limitation of one of the most extensively studied approaches to tackling the long-standing sample compression conjecture (Warmuth, 2003). Concretely, the approach proposed by Floyd and Warmuth (1995) entails embedding any given VC class into an extremal class of a comparable dimension, and then applying an optimal sample compression scheme for extremal classes. However, our results imply that this strategy would in some cases result in a sample compression scheme at least exponentially larger than what is predicted by the sample compression conjecture. The above implications follow from a general result we prove: any extremal class with VC dimension d has dual VC dimension at most 2d + 1. This bound is exponentially smaller than the classical bound 2d+1-1 of Assouad (1983), which applies to general concept classes (and is known to be unimprovable for some classes). We in fact prove a stronger result, establishing that 2d + 1 upper bounds the dual Radon number of extremal classes. This theorem represents an abstraction of the classical Radon theorem for convex sets, extending its applicability to a wider combinatorial framework, without relying on the specifics of Euclidean convexity. The proof utilizes the topological method and is primarily based on variants of the Topological Radon Theorem (Bajmóczy and Bárány, 1979). © 2024 Z. Chase, B. Chornomaz, S. Hanneke, S. Moran & A. Yehudayoff.

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：