摘要
我国的股票市场，从沪、深两个交易所成立日算起，已有十多年的历史。虽然中国股票市场具有与生俱来的制度性缺陷，但我们不能否定股票市场对中国经济增长的积极作用，它的跨越式的发展为国有企业改革和国民经济的持续健康发展做出了积极的贡献。
随着证券市场监管力度的加大以及扩容速度的增加，中国的证券市场会越来越规范，人为操纵价格的现象会大幅减少，对于股票市场的分析变得越来越重要。但目前流行的股票软件都是以简单的技术分析为主，进行价格预测的较少，使用了神经网络的就更少了。我们所研发的系统就是一个既能对股票进行常规的技术分析又能利用先进的神经网络方法进行预测的软件，并且预测的是股票价格而不是趋势。
二十一世纪是一个信息化的网络时代，人们获得股票数据和进行股票操作主要是通过互联网进行的。在设计时，我们的系统考虑到了应该实现以资源共享为核心的全新的股票分析和预测方式，这必将在一定程度上冲击传统的股票市场，并成为新世纪的主流。
传统的股票分析软件，对初入股票市场的用户来说是不太适合的。因为软件要求用户自己有对所提供的数据进行预测的能力，并对各股市场行情有充分的了解，但这常常是不可能做到的。此外，考虑到当前的软件相应的术语比较多，不利于初入股票交易市场的人使用，在用户操作上也有一定的难度，这就要有一个智能化的操作简单的软件产品来帮助这些用户来完成他们的需要。所以我们就想开发一个比较适合初入股票市场者使用的软件，在使用上基于用户的操作方便为前提，并且便于一般人群的掌握，即使是不懂股票知识的人使用起来也没有任何困难。
我们考虑到系统实现的方便性，在系统设计上采用了层次架构的方式，在体系结构中采用了C/S结构，这样既考虑到用户使用的灵活性和可扩展性，又兼顾了系统的安全性、稳定性与可靠性。在功能模块中主要是按照服务器端和客户端分别实现不同的功能来完成的。下面就具体介绍一下本系统所实现的功能。
服务器端共有七个功能模块：
1、数据采集
接收由卫星传送来的数据信息，得知各股信息。
2、技术分析
对数据进行处理,得出股市动态。
3、预测
主要是根据前面数据情况，来预测出股票在以后的价格。
4、系统维护
此功能主要是为保证系统的正常运行。
5、备份
对股票信息进行备份，可以做为以后的参考。
6、帮助
为用户使用提供方便，有不了解的问题可以进行查询。
7、退出
指本系统需要结束时，退出本系统。
客户端有五个功能模块：
数据采集（从服务器端采集数据）
通过自己定义的接口，从服务器端采集数据
用户接口
接收用户的请求并把结果返回给用户，以及用户间的交流
3、备份
用户对股票信息进行备份，可以作为以后的参考。
4、帮助
为用户使用提供方便，有不了解的问题可以进行查询。
5、退出
指本系统需要结束时，退出本系统。
本系统主要的功能有二个方面：一是分析，二是预测。
作为股票的技术分析这一个主要模块，目前在国内流行的软件中，在技术实现上比较成熟，也被大多数人所接受，所以本系统也遵循了一般的人们习惯，没有给予很大变化，只是在界面上给人的感觉更方便，更便于操作，比较适合于初学者使用。
本系统的独特之处在于，在股票价格预测模块中采用了神经网络这种先进的智能计算工具，我们采用了RBF（Radial Basis Function）网络这种性能良好的前向网络,它具有很强的生物背景和逼近任意非线性函数的能力。与其它前向网络相比，RBF网络在结构上具有训练速度快速、易行的特点。
为检验本系统中预测模块的效果，我们选择了从01/04/1999到06/30/2000之间连续300多个交易日的数据。考虑到收盘价是当日行情的最后价格，又是下一个交易日开盘价的重要参考，可以用它来预测未来股票市场行情，所以我们采用了收盘价作为计算依据。我们所使用的股票数据是日收盘价，是上海股票交易所的综合指数（Composite Index in Shanghai Stock Exchange, CISSE）。通过这些真实的数据检验，证明了神经网络能够进行具有很高精度的预测。从而为客户更好地把握股价的波动提供了强有力的手段。
对于本系统的不足之处，有待于进一步改进。
More than ten years have been passed since the establishment of the Shanghai Stock Exchange and the Shenzhen Stock Exchange. Though has its innate system defect, Chinese stock market is positive function to the growth of Chinese economy. Its swift development has made positive contribution to the SOE reform and continuous development of macro economy.
With the increasing supervision to the securities market and the increasing dilatation speed, the Chinese stock market will be more and more standardized. The manual control to the stock price will be reduced remarkably. The analysis of the stock market has become more and more important. But the popular stock software at present almost relies on simple technology analyses. The price forecasting is carried on scarcely, specially the use of neural networks. The software that we research and develop can carry on not only routine technology but also stock price forecasting with the advanced neural networks. Furthermore, it can forecast the stock price but the stock trend.
In the 21st century, an informative network era, the acquiring of stock data and the selling and buying of the stock can be done through the internet. So in our designing, a brand-new way for stock analysis and forecasting have been considered in our software which will shock the traditional stock market in some extent, and will become the mainstream in the new century.
Traditional stock analysing software isn’t suitable for the beginner of the market. Because the software requires the user must have the ability of forecasting using the data the software offered and the user must have an abundant understanding of the stock market. Both of the above can’t be accomplished easily. Moreover there are many difficulty professional terms in the present software which will not benefit to the user who has no in-depth understanding to the stock market. The operation of the software is also difficult to the user in some extent. So we want to develop relatively suitable software to help the new shareholder of the stock market. Our system is fit to the beginner. The convenience of operating is the most important character of the system. Even those who know nothing about the stock can operate it without difficulty.
Considering the design convenience of the system, we adopted the way in which the layer in the design of the system. C/S is adopted in the system structure. In the design of the system, the flexibility, extendibility, safety, stability and reliability of system is considered. Different function can be realized by server and client module. The function in our system will be introduced below.
Seven functional modules in the server:
1、 Data gathering ( gather the data from the server)
Through the interface that is defined by system, gather the data from the server.
2、Technical analyzing
Deal with the data to get the trend of the stock market.
3、Forecast
According to the data gathered to forecast the stock price in the future.
4、Maintenance of the system
The function is used to guarantee the normal operation of the system.
5、Backup
Put the stock information into file for later reference.
6、Help
Supply convenience for the user, to inquire when having some unknown questions.
7、Exit
To exit when the system is over.
Five functional modules in the client:
1、Data gathering
To know each stock information by receiving the data information sent by the artificial satellite.
2、User interface
Receive the request of user， return answer to user and communicate experience.
3、Backup
Put the stock information into file for later reference.
4、Help
Supply convenience for the user, to inquire when having some unknown questions.
5、Exit
To exit when the operation is over.
There are two dominating function in the system, one is analyzing, the other is forecasting.
Analyzing in popular software at present is relatively ripe which is accepted widely. So our system follows general habits of user with no great change in it. The interface is more convenient, easily operated, relative
引文
Cao L J and Tay Francis E H . Financial forecasting using support vector machines. Neural Computing & Applications, 2001, 10(2): 184-192.
Tay F E H and CAO L J. Application of support vector machines in financial time series forecasting [J]. The International Journal of Management Science, 2001, 29 (4): 309-317.
Tay F E H and Cao L J. ε－Descending support vector machines for financial time series forecasting [J]. Neural Processing Letters, 2002,15(2): 179-195
4．杨朝军。证券投资分析。上海人民出版社，2002。
5. 郭棣。基于RBF神经网络的软仪表的开发及在蒸馏装置的应用。微计算机信息，2003，19（8）：51－52。
6. 姜静清，宋初一，刘娜仁，张品一。神经网络的训练方法及分析。内蒙古大学学报（自然科学版），2003，18（4）：301-303。
7. 孙延风。吉林大学博士学位论文，2003。
