| ＲＢＦ神经网络在量化投资中的应用 |
| Alternative Title | RBF neural network in quantitative investment
|
| 周骐 |
| Thesis Advisor | 严定琪
|
| 2013-05-16
|
| Degree Grantor | 兰州大学
|
| Place of Conferral | 兰州
|
| Degree Name | 学士
|
| Keyword | ＲＢＦ神经网络
Ｇａｕｓｓｉａｎ函数
归一化算法
滚动预测
|
| Abstract | RBF 神经网络的研究是人工智能领域极具挑战性的研究之一，它可以解决优
化智能计算、智能模式控制以及模式识别等许多问题，RBF 神经网络在量化投资
中的应用也有很大的发展空间，主要是实现股票函数的趋势拟合。
本文利用Gaussian 函数作为径向基函数，构建具有输入层、隐含层和输出
层的神经网络结构，对“航天电子(600879)”股票数据进行拟合，选用Gaussian
型RBF 神经网络，利用具有单隐层n+1 个神经元的网络模型精确插值n+1 个样本，
根据样本值构造出Gaussian 型RBF 神经网络的内部和外部权值，从而达到预测
的目的。
首先我们对不同属性的股票数据进行归一化预处理，得到统一格式的数据，
便于处理。其次采用滚动预测这种迭代式算法构建预测框架，最后通过matlab
神经网络工具箱，对RBF 神经网络的输出值进行误差分析：分别计算出均值误差、
均方误差，百分比误差三种不同的误差进行全面分析。然后对比预测数据和测试
数据分析完善并推广模型。 |
| Other Abstract | RBF neural network is a challenging field of artificial intelligence research, it can solve the
optimization of intelligent computing, intelligent mode control, and pattern recognition and many
other issues, RBF neural network applications have much room for development in the
quantitative investment, mainly to achieve a stock function fitting.
Using Gaussian function as the radial basis function to construct the input layer, hidden layer and
output layer of the neural network structure, "Aerospace Electronics (600879)" stock data fitting,
the choice of a Gaussian RBF neural network, the use of single hidden layer neurons (n +1)
network model is accurate interpolation for n +1 samples, internal and external weights of the
Gaussian RBF neural network is constructed based on the sample values , so as to achieve the
purpose of the forecast.
First, we normalized pretreatment stock data of the different attributes, unified data format, easy to
deal with. Secondly, rolling forecasts of the iterative algorithm to construct the framework of
projections, and finally, the output value of the RBF neural network matlab neural network
toolbox Error Analysis: calculate the mean error, mean square error, percentage error of three
different error a comprehensive analysis. And then compare the forecast data and test data analysis
to improve and promote the model. |
| URL | 查看原文
|
| Language | 中文
|
| Document Type | 学位论文
|
| Identifier | https://ir.lzu.edu.cn/handle/262010/225664
|
| Collection | 数学与统计学院
|
Recommended Citation
GB/T 7714 |
周骐. ＲＢＦ神经网络在量化投资中的应用[D]. 兰州. 兰州大学,2013.
|
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
