决策树在数据挖掘中的新进展和发展前景

决策树是数据挖掘分类方法的一种。本文简单介绍了决策树及其生成过程和算法；着重叙述了决策树近年来在数据挖掘中的主要进展，探讨了各个方面的优缺点；讨论了目前决策树技术面临的挑战和发展前景。     

PCC系统构筑智能移动宽带

通过对PCC架构、PCC业务实现原理的分析,结合中国联通福建分公司PCC建设实例,探讨PCC架构如何构建智能移动宽带,以实现精细化管控,在提高运营商收益的同时,提升用户的网络体验。     

INDIVIDUAL COMMUNICATION TRANSMITTER IDENTIFICATION BASED ON MULTIFRACTAL ANALYSIS

In this letter, the communication transmitter transient signals are analyzed based on the time-variant hierarchy exponents of multifractal analysis. The species of optimized sample set is selected as the template of transmitter identification, so that the individual communication transmitter identification can be realized. The turn-on signals of four transmitters are used in the simulation. The experimental results show that the multifractal character of transmitter transient signals is an effective character of individual transmitter identification.     

基于图像退化模型的红外与可见光图像融合方法

基于深度学习的红外与可见光图像融合算法依赖人工设计的相似度函数衡量输入与输出的相似度,这种无监督学习方式不能有效利用神经网络提取深层特征的能力,导致融合结果不理想。针对该问题,该文首先提出一种新的红外与可见光图像融合退化模型,把红外和可见光图像视为理想融合图像通过不同退化过程后产生的退化图像。其次,提出模拟图像退化的数据增强方案,采用高清数据集生成大量模拟退化图像供训练网络。最后,基于提出的退化模型设计了简单高效的端到端网络模型及其网络训练框架。实验结果表明,该文所提方法不仅拥有良好视觉效果和性能指标,还能有效地抑制光照、烟雾和噪声等干扰。     

大数据标准化现状及展望

首先介绍大数据标准化的发展背景和面临的挑战,概述国内外大数据标准化发展现状,进一步详细介绍美国国家标准与技术研究院在大数据标准化方面的工作进展,最后对我国大数据标准化的工作方法和标准化体系提出建议。     

一种条件概率与遗传算法相结合的分类方法

数据挖掘就是从大量数据中发现以前未知的有用信息、模式、趋势的过程。分类是数据挖掘的一种主要方法。文章指出分类的实质是找出各属性对分类的贡献大小，然后采用分而治之的思想，先用条件概率的方法计算单个属性对分类的贡献，再利用遗传算法计算各属性对分类的重要程度，提出了条件概率与遗传算法相结合的分类方法，利用UCI数据集进行验证，并与相同条件下的其它分类方法进行了比较，实验表明该方法是一种简单有效的分类方法。     

基于影响函数的k-近邻分类

分类是一种监督学习方法,通过在训练数据集学习模型判定未知样本的类标号。与传统的分类思想不同,该文从影响函数的角度理解分类,即从训练样本集对未知样本的影响来判定未知样本的类标号。首先介绍基于影响函数分类的思想;其次给出影响函数的定义,设计3种影响函数;最后基于这3种影响函数,提出基于影响函数的k-近邻(kNN)分类方法。并将该方法应用到非平衡数据集分类中。在18个UCI数据集上的实验结果表明,基于影响函数的k-近邻分类方法的分类性能好于传统的k-近邻分类方法,且对非平衡数据集分类有效。     

人工智能视域下工业大模型优化的路径研究

人工智能技术的快速发展为工业领域带来了深刻变革,工业大模型作为新一代人工智能技术的重要载体,在推动工业智能化转型中发挥关键作用。针对工业大模型在实际应用中存在的数据质量参差、模型泛化能力不足、算力资源受限等问题,文中从数据层面、模型架构、计算优化和部署应用4个维度探讨了优化路径。通过分析工业场景特征,提出了基于领域知识的数据增强方法、轻量化模型裁剪策略、分布式训练加速方案和边缘智能部署框架,为提升工业大模型性能和应用效果提供了新思路。     

多播网络中基于网络编码的高效丢失恢复机制

网络编码为无线网络中可靠多播通信提供了有效解决途径。该文分析了网络中编码机会的变化规律,研究了解码失败的编码数据包对网络编码性能的影响,提出了新的基于网络编码的丢失恢复算法(NCLR)。NCLR要求节点缓存解码失败的编码数据包,并反馈信息给发送节点。根据各个节点的丢包情况,NCLR通过优先传输对编码性能影响较大的数据包,并在需要重传的已编码数据包和原始数据包中选择编码组合,来充分挖掘网络中的编码机会。仿真结果表明相对于已有算法,NCLR算法可以在重传次数和丢失恢复时延方面有显著性能改善。     

基于改进型深度数据流形的数据分类算法及在人脸中的应用

针对数据分类问题的局限,提出一种基于改进型深度数据流形的数据分类算法并将其应用到人脸识别中。首先,通过采集人脸图像的深度信息,利用稀疏表示对其进行去噪处理;再结合图像的颜色信息,重新生成三维人脸信息数据库,通过对人脸数据的流形分析得到最优的降维结果,按十字十乘交叉验证法的原则选取训练集和测试集,将训练集输入支持向量机算法建立数据分类器;最后,将测试集输入训练完成的分类器中,实现人脸数据分类。选取ORL,Yale两类人脸图像标准数据库与传统人脸识别算法进行交叉对比实验,验证算法的优越性和可行性。实验结果表明：所提出的算法有较高的分类准确率,可有效地完成人脸识别。     

CB-D2RNet – An efficient context bridge network for glioma segmentation

The recent automatic glioma segmentation and localization techniques obtained promising results, but there is much scope for improvement in execution complexity and segmentation efficiency. These methods often fail to pinpoint small and isolated target locations in necrotic and enhancing glioma sub-regions. Moreover, the computational complexity and number of model parameters utilized in these techniques are also high. To address such issues, a Context Bridge-Dense Dilated Residual Net (CB-D²RNet) is proposed in this paper which reflects the five novel contributions. Firstly, a Dense Dilated Convolutional (DDC) block is formed with four cascade branches to cope with large morphological differences in gliomas. Secondly, the skip connections in traditional UNet are redesigned to overcome the large contextual gap between encoder-decoder. Thirdly, a new loss function is proposed that handles unequal class distribution in gliomas and provides a regularization impact. Fourthly, the precise selection of dilation rates is made for each dilated convolutional block in the feature encoder to gather a more receptive view of complex and multiple tumor regions. Lastly, only a single convolutional operation is included in the feature encoder and decoder, unlike other state-of-the-art models. The experiments are conducted on BraTS 2018 and BraTS 2019 benchmarks, demonstrating that the proposed model performs competitively in all three glioma sub-regions. It achieves dice similarity coefficient for the whole tumor, tumor core, and enhancing tumor as 0.982, 0.987, and 0.976, respectively for the BraTS 2018 dataset, whereas 0.983, 0.989, and 0.977 respectively for the BraTS 2019 dataset. Besides this, the model uses only 6.7 million parameters, the lowest among other compared models.     

一种基于学习的自适应哈希算法研究

通常在一般关系数据库中采用的哈希函数都是针对某一应用而设计的。在具体应用中该函数也许是最优化的．但不能保证该函数适用于其他应用场合。本文提出一种基于枚举的自适应哈希算法并对该算法进行研究。实验表明．该算法能够使数据分布达到最优化，显著地提高数据的存取和查询效率。     

Enhancing IoT anomaly detection performance for federated learning

Federated Learning (FL) with mobile computing and the Internet of Things (IoT) is an effective cooperative learning approach. However, several technical challenges still need to be addressed. For instance, dividing the training process among several devices may impact the performance of Machine Learning (ML) algorithms, often significantly degrading prediction accuracy compared to centralized learning. One of the primary reasons for such performance degradation is that each device can access only a small fraction of data (that it generates), which limits the efficacy of the local ML model constructed on that device. The performance degradation could be exacerbated when the participating devices produce different classes of events, which is known as the class balance problem. Moreover, if the participating devices are of different types, each device may never observe the same types of events, which leads to the device heterogeneity problem. In this study, we investigate how data augmentation can be applied to address these challenges and improving detection performance in an anomaly detection task using IoT datasets. Our extensive experimental results with three publicly accessible IoT datasets show the performance improvement of up to 22.9% with the approach of data augmentation, compared to the baseline (without relying on data augmentation). In particular, stratified random sampling and uniform random sampling show the best improvement in detection performance with only a modest increase in computation time, whereas the data augmentation scheme using Generative Adversarial Networks is the most time-consuming with limited performance benefits.     

A classification-based privacy-preserving decision-making for secure data sharing in Internet of Things assisted applications

The introduction of the Internet of Things (IoT) paradigm serves as pervasive resource access and sharing platform for different real-time applications. Decentralized resource availability, access, and allocation provide a better quality of user experience regardless of the application type and scenario. However, privacy remains an open issue in this ubiquitous sharing platform due to massive and replicated data availability. In this paper, privacy-preserving decision-making for the data-sharing scheme is introduced. This scheme is responsible for improving the security in data sharing without the impact of replicated resources on communicating users. In this scheme, classification learning is used for identifying replicas and accessing granted resources independently. Based on the trust score of the available resources, this classification is recurrently performed to improve the reliability of information sharing. The user-level decisions for information sharing and access are made using the classification of the resources at the time of availability. This proposed scheme is verified using the metrics access delay, success ratio, computation complexity, and sharing loss.