改进的指数双向联想记忆模型在数据压缩中的应用

Data Compression by Improved Exponential Bidirectional Associative Memory Model

改进的指数双向联想记忆模型 (Improved exponential bidirectional associative memorymodel,Ie BAM)是在 e BAM的基础上通过引入内连接项而产生的一个比 e BAM具有更高存储容量、纠错性的联想神经网络。借助于 Ie BAM的高存储容量和良好的纠错性能及有序直方图技术 ,可实现一种更高效率的数据压缩方法 ,从而为现有的数据压缩方法提供一种新算法。最后 ,计算机模拟证实了使用 Ie BAM的数据压缩性能比使用 e BAM的更好。

Improved exponential bidirectional associative memory model(IeBAM), which was developed from eBAM by adding an intraconnection to the exponents, is a neural network with higher storage capacity and correcting error capability than eBAM. Using IeBAM′s high capacity and correcting error capability and the ordered histogram technique, a data compression algorithm with better efficiency is achieved, and then a new data compression algorithm to existing data compression method is supplied. Finally, simulational results of computer verify that the effect of data compression with IeBAM is much better than that with eBAM.