用Javabean构建数据过滤的Script代码生成器

针对asp或jsp的客户端数据过滤程序代码由手工编写,编程效率低的缺点,提出了用脚本代码生成器实现自动生成数据过滤代码的方法。生成器的主要组件是Javabean,它可以连接到不同的数据库并操作数据表,在提取数据表的字段结构信息后,由用户选择需要过滤的字段,生成器能根据数据表的结构参数匹配bean中的相关过滤函数,最后生成具有数据过滤能力的客户端脚本代码文档。将生成器用于Web程序的开发,不仅能减少开发人员在编写、测试和修改代码工作量,而且能及时跟踪数据库(表)设计的改变,脚本代码生成高效、自动、准确。     

基于社交网络的推荐系统研究

近年来,基于社交网络的推荐系统随着社交媒体和大数据的蓬勃发展,逐渐成为推荐系统重点研究方向。将社交网络用户社会化属性信息和评论内容与深度学习等技术结合,可有效解决传统推荐系统数据稀疏和冷启动等问题。首先回顾传统推荐系统常用方法,介绍社交网络推荐系统主要流程和基本框架,然后介绍最新相关研究方向和应用情况,最后对基于社交网络的推荐系统发展趋势进行分析与展望。     

基于决策树和bloom filter的包过滤防火墙

分析了基于链表的包过滤防火墙的性能和特点,针对其不足,将决策树和Bloom Filter引入到包过滤防火墙中,并对其中涉及到的算法和步骤的开销进行了分析.分析表明,这种方法可以有效提高包过滤防火墙的性能.     

外语教学课堂中的游戏

语言习得的过程以及习得的效果在很大程度上会从习得者心理上，特别是情感上加强过滤作用，因此语言教学的过程应当最大化教学内容、教学方法的多样性来降低这种过滤作用对习得的迁移作用。语言教学课堂中的游戏绝不仅仅只是吸引学生注意力，打发教学任务完成时间的工具，而恰恰可以鼓励习得者在不知不觉之间真正地用语言交流。本文通过对一些常用的具有很高调适性的语言教学课堂中的游戏的总结，希望使语言课堂中的游戏在辅助教学的过程中成为语言教学以及教学者和习得者之间的纽带。     

基于包过滤的Windows 2000/XP个人防火墙的设计与实现

研究了在Windows 2000/XP操作系统下拦截网络数据包的各种技术,采用Winsock2 SPI与NDIS HOOK相结合的过滤技术,设计和实现了一个个人防火墙。     

基于神经网络的垃圾邮件过滤

垃圾邮件的过滤是一个具有重要现实意义的课题．将交叉覆盖学习算法和向量空间模型等技术相结合可得到一种新的垃圾邮件过滤方法．实验结果表明该方法识别率较高，具有较强的实用价值．     

Improving the precision of the keyword-matching pornographic text filtering method using a hybrid model

With the flooding of pornographic information on the Internet, how to keep people away from that offensive information is becoming one of the most important research areas in network information security. Some applications which can block or filter such information are used. Approaches in those systems can be roughly classified into two kinds:metadata based and content based. With the development of distributed technologies, content based filtering technologies will play a more and more important role in filtering systems. Keyword matching is a content based method used widely in harmful text filtering. Experiments to evaluate the recall and precision of the method showed that the precision of the method is not satisfactory, though the recall of the method is rather high. According to the results, a new pornographic text filtering model based on reconfirming is put forward. Experiments showed that the model is practical, has less loss of recall than the single keyword matching method, and has higher precision.     

一种有效缓解数据稀疏问题的协同过滤推荐算法

传统协同过滤推荐算法依据共同评分项目计算用户相似度, 进而产生推荐项目。然而, 随着用户和商品数量的不断增加, 用户共同评分的项目会越来越少, 甚至没有, 因此传统协同过滤推荐算法对用户之间相似度的衡量将会越来越不准确, 从而影响推荐系统的性能。针对这一问题, 本文对用户相似度的计算方法进行了改进, 提出直接相似度和间接相似度的概念, 同时引入关键人物权重, 进一步提高推荐系统的准确性。     

WPT-based modal control on distributed structures with MRF-04K damper

The magnetorheological (MR) fluid damper-based semiactive control systems have received considerable attention for protecting structures against natural hazards such as strong earthquakes and high winds. In this paper, a novel modal controller using wavelet packet transform (WPT) is proposed for the vibration control of distributed structures. In the proposed control system, the WPT method is utilized to decompose the acceleration measurement and select the modes containing most of the WPT energy component as the dominant modes. Then, a modal controller is designed to control the dominant modes and the optimal active control force is solved. Finally, Clipped-optimal con- trol law is adopted to determine the voltage applied to each MR damper. A Kalman-filter observer, which estimates the full controlled modal states from local accelerometer feedbacks, is designed for rendering the controller to be more applicable to distributed structures with a large number of degrees of freedom. A numerical example of a stadium roof structure installed with MRF-04K damper is presented. The effectiveness of the controller is evaluated under both Tianjin and El Centro earthquake excitations. The superior performance and adaptability of the controller for versatile loading conditions are demonstrated through the comparison with traditional truncated modal controller.     

基于图像处理的工件检测

研究了基于图像处理的工件轮廓检测技术的流程及算法,通过图像滤波、图像增强、阈值分割、边缘检测、边缘定位和参数测量及拟合等一系列算法,可以测出被检测工件是否符合要求及其误差范围。