Student practices are an important aspect in schooling. This article explores a possible solution of this issue for Web students in the field of computer science. Taking a DBMS implementation course as an example, the practical experiences are introduced, including courseware organization, question answering, project assignments and checking for Web Learning. The courseware is featured with: (a) self-navigation by semantic links, (b) seeing before learning, (c) provision of source code and practice-slots, (d) learning with tasks, and (e) personalized interface facilities. The special techniques inside the implementation of a personalized interface, such as user information table, behavior table, topic level tree and the page producing mechanisms are also discussed. A strategy to get the feedback of Web students is proposed, namely, by setting up a shadow class on campus to reflect the feedback of all Web learning students.

Time series prediction is a typical and signicant task in data mining being widely studied recently. This paper proposes two novel methods for Time Series Prediction based on GEP (Gene Expression Programming). The main contributions include: (1) GEP-Sliding Window Prediction Method (GEP-SWPM) to mine the relationship between future and historical data directly. (2) GEP-Differential Equation Prediction Method (GEP-DEPM) to mine ordinary differential equations from training data, and predict future trends based on specied initial conditions. (3) A brand new equation mining method, called Difierential by Microscope Interpolation (DMI) that boosts the e

在许多科学研究中，人们希望揭示隐藏在现象背后的规律，并用函数关系来表示。论文分析了函数关系表达式挖掘技术的特点，提出了一种基于基因表达式编程的函数关系发现方法。MEM方法能处理具有一致表达式的关系和具有不同分域表达式的复杂函数关系。论文对该方法的复杂度和性能做了评价，论证了MEM方法具有对数数量级的复杂度。实验结果显示，基于GEP的函数关系发现方法在采用较高变异概率时有很好的性能，对于不同的目标函数，挖掘成功率可以达到20%-80%，且运行时间较短，成功挖掘平均耗时在10秒以内。

利用GEP方法实现了多项式函数分解。主要工作如下：提出了基于基因表达式编程GEP的多项式函数分解方法GPF（GEP Dnlvnomial factorization)。能将任意多项式函数关系，按指定精度分解若干个低次多项式函数之积。利用基因表达式编程GEP技术提出了GPF算法。采用了有特色的概率相关因子优化对GEP中的适应度函数进行优化，使得精度提高了27％。提出了宽松环境进化策略LEE（loose environment evnlution），使GEP成功率比传统技术最大提高了58倍。通过一系列的实验证明了GPF的有效性。

用传统基因表达式编程（GEP）适应度机制挖掘函数关系容易受到噪声干扰，导致结果失真，为此做了如下探索：①借鉴生物具有的“趋利避害”天性，提出了GEP的“弱适应模型”，以实现在含噪声的数据集上挖掘函数关系；②提出新概念“带内集”、“带外集”并用于划分训练数据集；③设计了在弱适应模型下基于相对误差计算适应度的算法REFA；④用详尽的实验验证了REFA的有效性，当测量数据的噪声率为3.33%时，与传统方法相比，REFA方法的成功率提高了3倍，产生结果的平均相对误差从7.899%降低到2.320%.