人工神经网络BP算法密集型数据的预处理

A Preprocessing Method on Dense Type of Data Using Artificial Neural Network BP Algorithm

针对误差反向传播算法密集型数据输入问题，提出一种数据预处理方法。模拟实验表明，采用这种预处理技术可以极大地改善密集型数据输入时网络的学习效率。

A preprocessing procedure on dense type of data is presented using the error backpropaga- tion algoritlim. Simulated results show that the learning effectiveness of the networks can be im-proved dramatically by employing the preprocessing technique as the input vectors are dense type ofdata.