数据挖掘是近年来计算科学新兴的一个研究方向。其主要研究目的是从大量数据中提取有用的信息，挖掘出潜在的规则，发现隐藏在数据中的规律。 将数据挖掘研究得到的理论结果和具体算法应用于实践，是研究领域重要的环节。本文描述了 一个数据挖掘平台的设计实现过程，该系统基于网上问卷调查模型，其算法基础源于JAVA 开源软件 WEKA 中的源代码包。 论文首先给出系统设计的整体思路，解决开发之前的高风险的关键问题。然后对系统的开发过程和思路进行详细的阐述。最后在开发完成后，进一步总结个人开发小型软件工程项目的经验，从中找出一定的规律，为以后的开发提供流程上的借鉴，从而降低开发的风险。 在理论上，将数据挖掘方法和应用模型结合，为数据挖掘技术的应用提供新的思路和方法。同时，探索针对小型软件工程项目的降低开发风险、提高开发速度和效率的新方法。 在实践上，为新的数据挖掘法的测试提供了一个简单易用的平台，能够使研究者在系统中方便的嵌入算法代码。同时，也为基于其他模型的类似平台的实现提供了参考和借鉴。

Data Mining is a new research direction in computer science coming up recently and has become an attractive topic to many researchers. The main purpose of the research is to obtain more information from large amount of data, in another word, to explore latent rules hidden in the data which people interested in and useful to the future decision. Applying the theories and algorithms of data mining plays an import part of research work. This thesis describes the designing procedure of a web application platform intended for online data mining, based on the online survey model. The data mining related algorithm codes in the program come from the source package of a JAVA open source application called WEKA. In the thesis, we firstly give a brief introduction of the design thought, including the way to avoid the technical risk before designing work, and solve the problems key to the project.Then we introduce the course of designing work and finally sum up some experience during the course of development Theoretically, it tries to find a new way to apply the theories of data mining to a special model like online survey. Meanwhile, it explores some new methods on reducing exploitation risk and enhances exploitation speed and efficiency of small scale web application project. Practically, the designed platform fabricates a framework to extract useful information from the online survey data, and the user may get a deeper insight of the data through analysis done automatically by the program.In addition, it provides an easy way for developers to add new algorithms to the platform, by which they could test their codes and get a brief view of the performance of their new algorithms. It is also used for reference to the implementation of similar platforms based on other models.