基于正则前馈神经网络的股票时间序列数据库的知识发现

Knowledge Discovery in Stock Market Time Series Based on Regularized Feedforward Neural Networks

将正则最小二乘前馈网络学习算法应用于时间序列的知识发现。正则最小二乘算法将正则化网络和节点删除算法结合起来,大大提高了前馈网络的泛化性能。将其应用于股票时间序列数据库的暂态规则的知识发现,发现过程包括时间序列数据库预处理和数据挖掘(规则发现)两部分，实验结果表明预测效果良好。

This paper uses the learning algorithm of feedforward neural networks based on the regularized least squares on the knowledge discovery on time series databases. The algorithm improves the generalization performance of feedforward neural networks through combining the regularization and pruning technology. It demonstrates the method on the temporal rule discovery of stock market time series database. The process of knowledge discovery includes preprocessing of time series data and data mining (rule discovery). The experiment demonstrates the effectiveness of the algorithm.