项目子相似度融合的协同过滤推荐算法

针对用户评分数据稀疏性和项目最近邻寻找的不准确性问题,提出了一种项目子相似度融合的协同过滤推荐算法.该算法根据目标用户每一属性取值,选取与该属性值一致的用户作为用户子空间,并在此空间上计算目标项目与其他项目之间的相似度(称其为项目子相似度).在此基础上,以项目子相似度为依据选取目标项目的K最近邻,计算其预测评分;最后对用户不同属性上的预测评分进行加权求和,得到目标项目的最终评分.实验结果表明,该算法能准确地选取目标项目的最近邻,明显改善了推荐质量.     

Inductive transfer with context-sensitive neural networks

Context-sensitive Multiple Task Learning, or csMTL, is presented as a method of inductive transfer which uses a single output neural network and additional contextual inputs for learning multiple tasks. Motivated by problems with the application of MTL networks to machine lifelong learning systems, csMTL encoding of multiple task examples was developed and found to improve predictive performance. As evidence, the csMTL method is tested on seven task domains and shown to produce hypotheses for primary tasks that are often better than standard MTL hypotheses when learning in the presence of related and unrelated tasks. We argue that the reason for this performance improvement is a reduction in the number of effective free parameters in the csMTL network brought about by the shared output node and weight update constraints due to the context inputs. An examination of IDT and SVM models developed from csMTL encoded data provides initial evidence that this improvement is not shared across all machine learning models.     

Dynamical characteristics of software trustworthiness and their evolutionary complexity

Developing trusted softwares has become an important trend and a natural choice in the development of software technology and applications, and software trustworthiness modeling has become a prerequisite and necessary means. To discuss and explain the basic scientific problems in software trustworthiness and to establish theoretical foundations for software trustworthiness measurement, combining the ideas of dynamical system study, this paper studies evolutionary laws of software trustworthiness and the dynamical mechanism under the effect of various internal and external factors, and proposes dynamical models for software trustworthiness, thus, software trustworthiness can be considered as the statistical characteristics of behaviors of software systems in the dynamical and open environment. By analyzing two simple examples, the paper explains the relationship between the limit evolutionary behaviors of software trustworthiness attributes and dynamical system characteristics, and interprets the dynamical characteristics of software trustworthiness and their evolutionary complexity. Supported partially by the National Basic Research Program of China (Grant No. 2005CB321900) and the National Natural Science Foundation of China (Grant No. 60473091)     

基于矩阵输入的ERP系统订单改进方案

本文在传统ERP系统的基础上,针对服饰、鞋帽、拉链等具有多属性产品的制造行业,提出一种运用矩阵输入对销售订单的改进方案。它从订单的特点出发,建立一般订单模型,详细介绍了矩阵式订单的设计,实现方法,论述了其适用范围和使用特点,最后分析了矩阵式录入订单的执行效率和可扩展性。     

汽车造型的工程属性与情感属性的映射关系研究

目的研究汽车造型中的工程属性与情感属性存在的映射和对应关系,以及对于设计评价的影响。方法通过文献研究和实验分析,揭示了汽车造型中的工程属性与情感属性的映射和转换关系。结论造型中的工程属性能与情感属性存在内在的映射与转化关系,可以提高设计方案在跨学科研发团队中的沟通效率,提高设计创意的"编码"和"解码"。对于汽车创意和评价具有重要的参考价值,可以提高设计方案在跨学科研发团队中的多向沟通效率,提高设计方案的接受度和可行性。     

面向海量NetFlow数据的存储和查询处理方法研究

针对全国骨干网高速海量Net Flow数据到来速度快、数据量大以及对所存数据进行频繁多维查询操作的特点,提出了一种多维属性聚簇存储(MACS)模型。该模型根据实际应用环境中查询的特点对数据进行空间分片,以并行加流水的方式对数据进行存储。此外,为Net Flow提出了一种超多面体的查询模式。真实环境实验结果表明,运用MACS模型实现的系统单点数据实时存储速度达到270万条/s,远远快于其他的数据分析系统,并且多维属性查询的速度优于Hive和Impala。     

基于Webdecoy的攻击行为特征提取研究

为更好地解决Web应用的安全问题,提出了基于Webdecoy的攻击行为特征提取模型.以具体的业务系统为基础,通过嵌入监控代码跟踪对业务系统的访问并获取探测攻击信息,并对获取信息进行归一化处理和属性划分,根据属性的动态比较得到攻击请求的相似性结构.最后对探测行为进行检测评估,对恶意行为给出报警,并通过实验结果验证了该模型的有效性.     

对设计做为生产力要素的思考

中国在改革开放以后,由"科技生产力"带领"中国制造"走向了世界。如今,在后工业时代背景下,对传统生产力重新反思并提出"设计生产力"的构思愈来愈显重要。笔者对设计作为生产力要素的多重属性和运行机制进行探索,并对设计生产力的社会意义进行归纳。     

Should we try to measure software quality attributes directly?

Most external software quality attributes are conceptually subjective. For example, maintainability is an external software quality attribute, and it is subjective because interpersonally agreed definitions for the attribute include the phrase ‘the ease with which maintenance tasks can be performed’. Subjectivity clearly makes measurement of the attributes and validation of prediction systems for the attributes problematic. In fact, in spite of the definitions, few statistically valid attempts at determining the predictive capability of prediction systems for external quality attributes have been published. When validations have been attempted, one approach used is to ask experts to indicate if the values provided by the prediction system informally agree with the experts’ intuition. These attempts are undertaken without determining, independently of the prediction system, whether the experts are capable of direct consistent measurement of the attribute. Hence, a statistically valid and unbiased estimate of the predictive capability of the prediction system cannot be obtained (because the experts’ measurement process is not independent of the prediction system’s values). In this paper, it is argued that the problem of subjective measurement of quality attributes should not be ignored if quality is to be introduced into software in a controlled way. Further, it is argued that direct measurement of quality attributes should be encouraged and that in fact such measurement can be quantified to establish consistency using an existing approach. However, the approach needs to be made more accessible to promote its use. In so doing, it would be possible to decide whether consistent independent estimates of the true values of software quality attributes can be assigned and prediction systems for quality attributes developed.