面向对象的神经网络类库

提出一种面向对象的神经网络类库的构造方法。该类库采用面向对象的方法，集成了当前比较成熟的人工神经网络模型，充分利用了ＯＯＰ方法的继承性、封装性和多态性，便于在现有基础上派生出新的神经网络模型。该类库将被用于实现一种最新的同步振荡神经网络－ＬＥＧＩＯＮ网络。由于系统的结构性和代码的可重用性，既有利于科研工作者进行神经网络的研究，也便于工程人员的神经网络应用。

The Neural Network Library Constructed Using OOP Method

A neural network library using object oriented programming method (OOP) is built. The library takes advantage of the inheritance, capsulation and multi status character of the OOP method and integrates almost all of the popular neural network model. Because of the structure and code reuse character of the library system, new neural network model can be easily integrated into the library, which is convenient not only for the neural network research, but also for the applications of the neural network in engineering area. A new neural network LEGION will be constructed using this library.